dj_pirtu wrote on 2022-04-06, 08:30:

Tried so far: Geforce3 Ti200 Geforce4 Ti4200 Quadro4 980XGL Geforce FX5900XT Geforce 6600GT Geforce 6800XT […]
Show full quote

Tried so far:
Geforce3 Ti200
Geforce4 Ti4200
Quadro4 980XGL
Geforce FX5900XT
Geforce 6600GT
Geforce 6800XT

None of those aren't fully stable. In FlatOut2 it's crashing in minutes or when idling in desktop. Sometimes blue screen, d3d9.dll error, sometimes screen freeze or reboots.
Radeon 9600, 9700, X800Pro are all fine, not one problem.

Tested so far a GeForce2 Ultra and a GeForce 6800GS, both pass a cycle of 3DMark2001SE (edit: 6800GS looped 3DMark for a couple of hours since I posted this, no issues).

BTW, took my Abit BF6 with 1,4GHz Tualatin out. Runs @148FSB so 1,55GHz, Geforce 5900XT, Voodoo2, Vortex2, Audigy2, ESS SOLO-1 and Gravis Ultrasound. Not a single problem. Intel BX forever.

I'll have to run some more tests... I want to re-do the testing, but this time I will confirm system stability beforehand with something like Linpack because Prime95 doesn't always catch everything. I often find myself confounded when I get mixed results only to find that such and such board doesn't like such and such RAM stick. So far, yes, I suspect the KT7A has an issue with AGP, but I am also partially convinced that externally powered cards should work ok. Yes, I think your board is defective as well in some other manner.

For what it's worth. I too had some weird issue with Abit KT7A V1.3, that I would not get a picture with Geforce 4 MX 440-SE, but on another same exact KT133A chipset based motherboard QDI Kinetiz-7E, there was no problems...

Have to test Geforce FX-5900 Ultra and 6800 XT on the KT7A later. Those cards also had zero issues on the Kinetiz-7E. If there are going to be any issues, the only significant difference I see is the linear regulator related to the AGP slot, which only powers the logic side, I would guess, and not the GPU / graphics RAM, for which power is typically regulated on the AGP card itself from 5 or 12 volts.

makechu wrote on 2023-10-23, 20:17:

For what it's worth. I too had some weird issue with Abit KT7A V1.3, that I would not get a picture with Geforce 4 MX 440-SE, but on another same exact KT133A chipset based motherboard QDI Kinetiz-7E, there was no problems...

Have to test Geforce FX-5900 Ultra and 6800 XT on the KT7A later. Those cards also had zero issues on the Kinetiz-7E. If there are going to be any issues, the only significant difference I see is the linear regulator related to the AGP slot, which only powers the logic side, I would guess, and not the GPU / graphics RAM, for which power is typically regulated on the AGP card itself from 5 or 12 volts.

If anyone can figure it out, I'm certain it is you... Perhaps there is a vdroop caused by a component with an incorrect value?

mockingbird wrote on 2023-10-23, 21:23:

makechu wrote on 2023-10-23, 20:17:

For what it's worth. I too had some weird issue with Abit KT7A V1.3, that I would not get a picture with Geforce 4 MX 440-SE, but on another same exact KT133A chipset based motherboard QDI Kinetiz-7E, there was no problems...

Have to test Geforce FX-5900 Ultra and 6800 XT on the KT7A later. Those cards also had zero issues on the Kinetiz-7E. If there are going to be any issues, the only significant difference I see is the linear regulator related to the AGP slot, which only powers the logic side, I would guess, and not the GPU / graphics RAM, for which power is typically regulated on the AGP card itself from 5 or 12 volts.

If anyone can figure it out, I'm certain it is you... Perhaps there is a vdroop caused by a component with an incorrect value?

Cap plague around that time is my guess. Recap the whole board

gmaverick2k wrote on 2023-10-25, 21:59:

Cap plague around that time is my guess. Recap the whole board

No, the board is re-capped. Entirely.

Try with another GPU similar spec. SMT component may have been knocked off

gmaverick2k wrote on 2023-10-25, 22:46:

Try with another GPU similar spec. SMT component may have been knocked off

Got several socket A boards, different brands, same chipset... GPUs are ruled out as the issue, but I will do more testing to verify again.

mockingbird wrote on 2023-10-23, 01:47:

dj_pirtu wrote on 2022-04-06, 08:30:

Tried so far: Geforce3 Ti200 Geforce4 Ti4200 Quadro4 980XGL Geforce FX5900XT Geforce 6600GT Geforce 6800XT […]
Show full quote

Tried so far:
Geforce3 Ti200
Geforce4 Ti4200
Quadro4 980XGL
Geforce FX5900XT
Geforce 6600GT
Geforce 6800XT

None of those aren't fully stable. In FlatOut2 it's crashing in minutes or when idling in desktop. Sometimes blue screen, d3d9.dll error, sometimes screen freeze or reboots.
Radeon 9600, 9700, X800Pro are all fine, not one problem.

Tested so far a GeForce2 Ultra and a GeForce 6800GS, both pass a cycle of 3DMark2001SE (edit: 6800GS looped 3DMark for a couple of hours since I posted this, no issues).

BTW, took my Abit BF6 with 1,4GHz Tualatin out. Runs @148FSB so 1,55GHz, Geforce 5900XT, Voodoo2, Vortex2, Audigy2, ESS SOLO-1 and Gravis Ultrasound. Not a single problem. Intel BX forever.

I'll have to run some more tests... I want to re-do the testing, but this time I will confirm system stability beforehand with something like Linpack because Prime95 doesn't always catch everything. I often find myself confounded when I get mixed results only to find that such and such board doesn't like such and such RAM stick. So far, yes, I suspect the KT7A has an issue with AGP, but I am also partially convinced that externally powered cards should work ok. Yes, I think your board is defective as well in some other manner.

Actually I had these problems with all three of my KT7A boards, two are fully recapped and made no difference. One of those boards I sold away some time ago.

One thing I noticed with Geforce 6800GT with VIA 694T board (havent tried with KT7A yet), 67.77 geforce drivers are really unstable. Crashes really easy in WinXP when 71.84 drivers works with no problems. Need to do that driver testing again, can't remember did I test all the different driver versions with all of the boards...

v1.2 and v1.3 KT7A boards work really well as long as you use an ISA sound card. If you use a PCI sound card you'll want to read up on the 686B southbridge bug to mitigate that problem. However, the v1.0 KT7A gave me headaches with weird issues that you can read about here.

Things I've learned about the v1.0 KT7A when installing Win98SE:

Don't install more than 512mb of RAM.
Radeon 9600XT and 9800XT cards work just fine.
NVIDIA cards newer than GF4 don't work.

With 512MB of RAM installed, I was able to successfully run Win98SE with a GF3 Ti 500 and a GF4 ti4600-8X (aka, GF4 Ti4800). A 5900 Ultra and a 6800GT did not work.

Note: All of my KT7A boards have been recapped by Chris over at badcaps.net.

Here's some info about the 686B bug:
https://www.theregister.com/2001/04/12/dataco … s_via_chipsets/

I'm also attaching a few files which may help others with this sort of thing.

When I had a KT7A in 2001 I used a Philips Acoustic Edge (PSC706) sound card with it, and it worked perfectly. I had no idea about the issues with Creative cards and Via chipsets, it was the best sound card that Best Buy had in stock when I needed a sound card.

mockingbird wrote on 2023-10-23, 21:23:

makechu wrote on 2023-10-23, 20:17:

For what it's worth. I too had some weird issue with Abit KT7A V1.3, that I would not get a picture with Geforce 4 MX 440-SE, but on another same exact KT133A chipset based motherboard QDI Kinetiz-7E, there was no problems...

Have to test Geforce FX-5900 Ultra and 6800 XT on the KT7A later. Those cards also had zero issues on the Kinetiz-7E. If there are going to be any issues, the only significant difference I see is the linear regulator related to the AGP slot, which only powers the logic side, I would guess, and not the GPU / graphics RAM, for which power is typically regulated on the AGP card itself from 5 or 12 volts.

If anyone can figure it out, I'm certain it is you... Perhaps there is a vdroop caused by a component with an incorrect value?

Unfortunately, I am having too much fun with the already perfectly functioning QDI Kinetiz-7E, so much so that I might not be in any hurry to come back to fiddling with the KT7A. Kinetiz-7E uses the same chipset, also has an ISA slot, and at least I have had zero issues with Geforce 4 Ti 4400, FX 5900 Ultra, and geforce 6800 w. 256 MB on it. The 6800 is the best card for it now, as the FX 5900 Ultra is just too loud... Kinetiz also it seems to work well with setmul and other slowdown software, like the ACPI based Throttle, with a mobile Barton CPU.
Maybe I'll come back to KT7A sometime in the future, if I feel I need to build another DOS compatible gaming PC for some LAN multiplayer stuff.

The Kinetiz-7E is an excellent and underrated MB. I'm glad you're having fun with it 😀

makechu wrote on 2023-10-26, 16:30:

Unfortunately, I am having too much fun with the already perfectly functioning QDI Kinetiz-7E, so much so that I might not be in any hurry to come back to fiddling with the KT7A. Kinetiz-7E uses the same chipset, also has an ISA slot, and at least I have had zero issues with Geforce 4 Ti 4400, FX 5900 Ultra, and geforce 6800 w. 256 MB on it. The 6800 is the best card for it now, as the FX 5900 Ultra is just too loud... Kinetiz also it seems to work well with setmul and other slowdown software, like the ACPI based Throttle, with a mobile Barton CPU.
Maybe I'll come back to KT7A sometime in the future, if I feel I need to build another DOS compatible gaming PC for some LAN multiplayer stuff.

Did you have to patch the bios for SETMUL support, or does it work out of the box?

makechu wrote on 2023-10-19, 01:21:

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required sw […]
Show full quote

mockingbird wrote on 2023-10-19, 01:08:

How about a hint then...? Sure, I'm familiar with the usual warnings of using polymers in a circuit that's not designed for too […]
Show full quote

makechu wrote on 2023-10-19, 01:02:

I'll have to see if I dare to post how I determined everything when I did the recapping, as I do not actually want anyone to follow in my footsteps with this. Probably better that I just won't do it... Knowledge just increases the pain of the process, as then there is no option to be just satisfied that the system is working, but also knowing that it could somehow be even better. But maybe that is just me...

How about a hint then...? Sure, I'm familiar with the usual warnings of using polymers in a circuit that's not designed for too low an ESR. A good example is old PSUs where the ripple becomes unacceptable when polymers are placed in the secondary filtering circuit. But P4 and later VRMs are almost universally compatible with very low ESR polymers...

Early Athlon VRMs aren't though.

I have another KT7A I want to re-cap, I'd like to get your secret recipe... What was it, the inductors? Did you change the inductors? Couldn't be the MOSFETs, could it?

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required switching adjustments and the use of the external compensation loop is not even required. As an example, in the KT7A, the CPU regulator is more tolerable to using very low ESR capacitors, but the chipset / RAM uses an older regulator IC that requires feedback loop compensation. The CPU regulator uses some sort of a shunt resistor with additional internal logic to implement some sort of "droop" compensation, and I assume the internal logic probably averages the ripple so that the noise components of the signal won't affect the switching of the MOSFETs directly, as it would in older regulator IC designs. Anyways, it is quite late here so I will call it a night here...

Edit: Since I started correcting something from the text, I could give as a hint that I only changed the chipset / RAM regulator's compensation. One ceramic capacitor had to be switched in the loop. Also, only the 1st zero and pole of the filter was implemented in the circuit on the board, and therefore only that can and has to be considered. The main electrolytic capacitors that are connected to the chipset / RAM regulator's output have to be identified, for example with a multimeter, to determine which capacitors are connected to it. I can say that there are 7 electrolytic capacitors, for which the original parallel ESR had to be determined to be able to calculate the required change. By reverse engineering the feedback loop circuit, I determined it was implemented for ~33 mOhms of parallel ESR with the original set of capacitors. My new parallel ESR was significantly lower, and I had to change the capacitor value of C2 = BC13 on the board for the feedback loop to be more towards the correct. The exact values required for the compensation loop will be difficult to determine, as it would require considering any SMD ceramic capacitors in the output as well, which there are too many to start desoldering to determine what they are. However, just using the main capacitor ESRs, I got a regulation result that seemed rock solid on the oscilloscope.
But maybe I said too much already, as like stated, I really don't want anyone to start messing with the tiny compensation loop components. Doing an error there might be able to actually destroy things on the board, and in the expansion slots....

Another edit... Correction to the value of the parallel ESR for the original KT7A chipset / RAM capacitor group; Actually, I estimated the average ESR per electrolytic capacitor to be about 25 - 33 mOhms originally, and the parallel ESR for the whole chipset / RAM capacitor group was only about 3.6 - 3.7 mOhms, which I determined by reverse engineering the feedback compensation circuit.
Well, as an excuce, my memory apparently failed when trying to do a quick reply in the middle of the night...

What about your QDI Kinetiz 7E? You said you recapped it all with polymer capacitors. Did you had to change anything else to make sure the external compensation loop was fine for the board to work normally? Or was it unnecessary because Kinetiz 7E has better ICs/voltage regulators than KT7A?

Windows98_guy wrote on 2023-11-16, 10:11:

makechu wrote on 2023-10-19, 01:21:

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required sw […]
Show full quote

mockingbird wrote on 2023-10-19, 01:08:

How about a hint then...? Sure, I'm familiar with the usual warnings of using polymers in a circuit that's not designed for too low an ESR. A good example is old PSUs where the ripple becomes unacceptable when polymers are placed in the secondary filtering circuit. But P4 and later VRMs are almost universally compatible with very low ESR polymers...

Early Athlon VRMs aren't though.

I have another KT7A I want to re-cap, I'd like to get your secret recipe... What was it, the inductors? Did you change the inductors? Couldn't be the MOSFETs, could it?

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required switching adjustments and the use of the external compensation loop is not even required. As an example, in the KT7A, the CPU regulator is more tolerable to using very low ESR capacitors, but the chipset / RAM uses an older regulator IC that requires feedback loop compensation. The CPU regulator uses some sort of a shunt resistor with additional internal logic to implement some sort of "droop" compensation, and I assume the internal logic probably averages the ripple so that the noise components of the signal won't affect the switching of the MOSFETs directly, as it would in older regulator IC designs. Anyways, it is quite late here so I will call it a night here...

Edit: Since I started correcting something from the text, I could give as a hint that I only changed the chipset / RAM regulator's compensation. One ceramic capacitor had to be switched in the loop. Also, only the 1st zero and pole of the filter was implemented in the circuit on the board, and therefore only that can and has to be considered. The main electrolytic capacitors that are connected to the chipset / RAM regulator's output have to be identified, for example with a multimeter, to determine which capacitors are connected to it. I can say that there are 7 electrolytic capacitors, for which the original parallel ESR had to be determined to be able to calculate the required change. By reverse engineering the feedback loop circuit, I determined it was implemented for ~33 mOhms of parallel ESR with the original set of capacitors. My new parallel ESR was significantly lower, and I had to change the capacitor value of C2 = BC13 on the board for the feedback loop to be more towards the correct. The exact values required for the compensation loop will be difficult to determine, as it would require considering any SMD ceramic capacitors in the output as well, which there are too many to start desoldering to determine what they are. However, just using the main capacitor ESRs, I got a regulation result that seemed rock solid on the oscilloscope.
But maybe I said too much already, as like stated, I really don't want anyone to start messing with the tiny compensation loop components. Doing an error there might be able to actually destroy things on the board, and in the expansion slots....

Another edit... Correction to the value of the parallel ESR for the original KT7A chipset / RAM capacitor group; Actually, I estimated the average ESR per electrolytic capacitor to be about 25 - 33 mOhms originally, and the parallel ESR for the whole chipset / RAM capacitor group was only about 3.6 - 3.7 mOhms, which I determined by reverse engineering the feedback compensation circuit.
Well, as an excuce, my memory apparently failed when trying to do a quick reply in the middle of the night...

What about your QDI Kinetiz 7E? You said you recapped it all with polymer capacitors. Did you had to change anything else to make sure the external compensation loop was fine for the board to work normally? Or was it unnecessary because Kinetiz 7E has better ICs/voltage regulators than KT7A?

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes about it now that I checked.
I believe it has a linear regulator for RAM / Chipset voltage, so there is nothing to adjust there, and most linear regulators are quite happy with very low ESR capacitors. I also checked the regulator circuit for the CPU, and if I remember correctly, it also uses droop-compensation similarly to KT7A, and has multiple output phases, so it seems to be fine with very low ESR capacitors and did not require any immediate change to any controller feedbacks. The CPU voltage did change slightly after recap; it was about 0.03 - 0.05 volt lower than before, if I recall correctly, but otherwise ok. But even with a little bit overclocked mobile Athlon XP CPU, I have had no issues with stability so I don't care about that.

But anyways, especially in the case of switched mode regulators, which are easily identifiable by having some inductors accompanied by some transistors, the output voltages must always be checked with an oscilloscope after recapping to see how the regulation is doing, and to see that there is no unexpected transient behavior when the system is stressed and the load levels are changing rapidly.

theseim wrote on 2023-11-15, 17:06:

makechu wrote on 2023-10-26, 16:30:

Unfortunately, I am having too much fun with the already perfectly functioning QDI Kinetiz-7E, so much so that I might not be in any hurry to come back to fiddling with the KT7A. Kinetiz-7E uses the same chipset, also has an ISA slot, and at least I have had zero issues with Geforce 4 Ti 4400, FX 5900 Ultra, and geforce 6800 w. 256 MB on it. The 6800 is the best card for it now, as the FX 5900 Ultra is just too loud... Kinetiz also it seems to work well with setmul and other slowdown software, like the ACPI based Throttle, with a mobile Barton CPU.
Maybe I'll come back to KT7A sometime in the future, if I feel I need to build another DOS compatible gaming PC for some LAN multiplayer stuff.

Did you have to patch the bios for SETMUL support, or does it work out of the box?

I had to do a patch with PCISET utility to enable the feature to change the mobile CPU multiplier in software. I guess it can be found here somewhere. Works even if the CPU is pinmodded, i.e. small wires put to the socket to change the CPU parameters. Off the top of my head, the bit to change was somewhere in the PCI register residing at address 55h. I just made a batch file to run the PCISET whenever I needed to test anything with changing the multiplier.
I tried to make this change in to the BIOS image file to be flashed to EEPROM, but somehow I could not figure out how to edit the checksum of the BIOS file without always having some error, so I just gave up on that idea. Should be doable though, if someone has the time.

makechu wrote on 2023-11-18, 20:53:

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes […]
Show full quote

Windows98_guy wrote on 2023-11-16, 10:11:

makechu wrote on 2023-10-19, 01:21:

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required sw […]
Show full quote

It has to be checked from the controller IC datasheets. Some newer regulator ICs have smarter methods of determining required switching adjustments and the use of the external compensation loop is not even required. As an example, in the KT7A, the CPU regulator is more tolerable to using very low ESR capacitors, but the chipset / RAM uses an older regulator IC that requires feedback loop compensation. The CPU regulator uses some sort of a shunt resistor with additional internal logic to implement some sort of "droop" compensation, and I assume the internal logic probably averages the ripple so that the noise components of the signal won't affect the switching of the MOSFETs directly, as it would in older regulator IC designs. Anyways, it is quite late here so I will call it a night here...

Edit: Since I started correcting something from the text, I could give as a hint that I only changed the chipset / RAM regulator's compensation. One ceramic capacitor had to be switched in the loop. Also, only the 1st zero and pole of the filter was implemented in the circuit on the board, and therefore only that can and has to be considered. The main electrolytic capacitors that are connected to the chipset / RAM regulator's output have to be identified, for example with a multimeter, to determine which capacitors are connected to it. I can say that there are 7 electrolytic capacitors, for which the original parallel ESR had to be determined to be able to calculate the required change. By reverse engineering the feedback loop circuit, I determined it was implemented for ~33 mOhms of parallel ESR with the original set of capacitors. My new parallel ESR was significantly lower, and I had to change the capacitor value of C2 = BC13 on the board for the feedback loop to be more towards the correct. The exact values required for the compensation loop will be difficult to determine, as it would require considering any SMD ceramic capacitors in the output as well, which there are too many to start desoldering to determine what they are. However, just using the main capacitor ESRs, I got a regulation result that seemed rock solid on the oscilloscope.
But maybe I said too much already, as like stated, I really don't want anyone to start messing with the tiny compensation loop components. Doing an error there might be able to actually destroy things on the board, and in the expansion slots....

Another edit... Correction to the value of the parallel ESR for the original KT7A chipset / RAM capacitor group; Actually, I estimated the average ESR per electrolytic capacitor to be about 25 - 33 mOhms originally, and the parallel ESR for the whole chipset / RAM capacitor group was only about 3.6 - 3.7 mOhms, which I determined by reverse engineering the feedback compensation circuit.
Well, as an excuce, my memory apparently failed when trying to do a quick reply in the middle of the night...

What about your QDI Kinetiz 7E? You said you recapped it all with polymer capacitors. Did you had to change anything else to make sure the external compensation loop was fine for the board to work normally? Or was it unnecessary because Kinetiz 7E has better ICs/voltage regulators than KT7A?

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes about it now that I checked.
I believe it has a linear regulator for RAM / Chipset voltage, so there is nothing to adjust there, and most linear regulators are quite happy with very low ESR capacitors. I also checked the regulator circuit for the CPU, and if I remember correctly, it also uses droop-compensation similarly to KT7A, and has multiple output phases, so it seems to be fine with very low ESR capacitors and did not require any immediate change to any controller feedbacks. The CPU voltage did change slightly after recap; it was about 0.03 - 0.05 volt lower than before, if I recall correctly, but otherwise ok. But even with a little bit overclocked mobile Athlon XP CPU, I have had no issues with stability so I don't care about that.

But anyways, especially in the case of switched mode regulators, which are easily identifiable by having some inductors accompanied by some transistors, the output voltages must always be checked with an oscilloscope after recapping to see how the regulation is doing, and to see that there is no unexpected transient behavior when the system is stressed and the load levels are changing rapidly.

I am looking at both kinetiz 7E and KT7A right now. I do see some inductors on the KT7A around the ram slots (2 I think) while the Kinetit 7e doesn't have them (it only has inductors at the cpu area just like KT7A).
Which component do you need to measure for the output voltage? The inductors or something else?

Windows98_guy wrote on 2023-11-19, 09:53:

makechu wrote on 2023-11-18, 20:53:

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes […]
Show full quote

Windows98_guy wrote on 2023-11-16, 10:11:

What about your QDI Kinetiz 7E? You said you recapped it all with polymer capacitors. Did you had to change anything else to make sure the external compensation loop was fine for the board to work normally? Or was it unnecessary because Kinetiz 7E has better ICs/voltage regulators than KT7A?

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes about it now that I checked.
I believe it has a linear regulator for RAM / Chipset voltage, so there is nothing to adjust there, and most linear regulators are quite happy with very low ESR capacitors. I also checked the regulator circuit for the CPU, and if I remember correctly, it also uses droop-compensation similarly to KT7A, and has multiple output phases, so it seems to be fine with very low ESR capacitors and did not require any immediate change to any controller feedbacks. The CPU voltage did change slightly after recap; it was about 0.03 - 0.05 volt lower than before, if I recall correctly, but otherwise ok. But even with a little bit overclocked mobile Athlon XP CPU, I have had no issues with stability so I don't care about that.

But anyways, especially in the case of switched mode regulators, which are easily identifiable by having some inductors accompanied by some transistors, the output voltages must always be checked with an oscilloscope after recapping to see how the regulation is doing, and to see that there is no unexpected transient behavior when the system is stressed and the load levels are changing rapidly.

I am looking at both kinetiz 7E and KT7A right now. I do see some inductors on the KT7A around the ram slots (2 I think) while the Kinetit 7e doesn't have them (it only has inductors at the cpu area just like KT7A).
Which component do you need to measure for the output voltage? The inductors or something else?

The best place to measure would be the voltages near the load. For example, quite often some unpopulated capacitor footprints connected to the relevant voltage rails can be found, which are good points to stick the oscilloscope probe. Note, that also the ground point should be selected to be as near to the measured load as possible, to be just measuring the voltage over the affected component and not some other noise, that can arise even from EMI affecting the probe's measurement leads themselves. Also, even if the ground plane is large and thick, with low resistance, this does not mean that it can be picked up from anywhere as there are plenty of voltage disturbances in the ground plane, which are not related directly to the voltage of the component being measured, which can give measurement results that do not correspond to what the target component is actually affected by. In a good measurement setup, there should not be a large loop area anywhere, and this includes what the probe leads form between + and - points.
I am not going to go in to basics here of what oscilloscope probes to use, or other oscilloscope basics, which I assume anyone playing with measuring power circuits with oscilloscopes should be aware of, but I typically use an isolated probe with 100 MHz bandwidth with short leads, which is more than enough for these kinds of measurements. I have also recently purchased a hand-held 100 MHz scope, and found that to be a good enough tool as well to measure voltages easily on motherboards, as there is no mains power grounding related issues to worry about with any probe.

The output side of the inductors could be measured also, if nothing else is easily accessible, but care has to be taken that the inductor EMI would affect the measurement probe as little as possible. The probe wires can easily pick up elevated levels of disturbance to the measurement due to the inductor's electromagnetic field, depending on how they are positioned in relation to the inductor.

makechu wrote on 2023-11-19, 17:53:

The best place to measure would be the voltages near the load. For example, quite often some unpopulated capacitor footprints co […]
Show full quote

Windows98_guy wrote on 2023-11-19, 09:53:

makechu wrote on 2023-11-18, 20:53:

It has been some time since I recapped the Kinetiz 7E, so my memories are becoming somewhat hazy of it, and I did not find notes about it now that I checked.
I believe it has a linear regulator for RAM / Chipset voltage, so there is nothing to adjust there, and most linear regulators are quite happy with very low ESR capacitors. I also checked the regulator circuit for the CPU, and if I remember correctly, it also uses droop-compensation similarly to KT7A, and has multiple output phases, so it seems to be fine with very low ESR capacitors and did not require any immediate change to any controller feedbacks. The CPU voltage did change slightly after recap; it was about 0.03 - 0.05 volt lower than before, if I recall correctly, but otherwise ok. But even with a little bit overclocked mobile Athlon XP CPU, I have had no issues with stability so I don't care about that.

But anyways, especially in the case of switched mode regulators, which are easily identifiable by having some inductors accompanied by some transistors, the output voltages must always be checked with an oscilloscope after recapping to see how the regulation is doing, and to see that there is no unexpected transient behavior when the system is stressed and the load levels are changing rapidly.

I am looking at both kinetiz 7E and KT7A right now. I do see some inductors on the KT7A around the ram slots (2 I think) while the Kinetit 7e doesn't have them (it only has inductors at the cpu area just like KT7A).
Which component do you need to measure for the output voltage? The inductors or something else?

The best place to measure would be the voltages near the load. For example, quite often some unpopulated capacitor footprints connected to the relevant voltage rails can be found, which are good points to stick the oscilloscope probe. Note, that also the ground point should be selected to be as near to the measured load as possible, to be just measuring the voltage over the affected component and not some other noise, that can arise even from EMI affecting the probe's measurement leads themselves. Also, even if the ground plane is large and thick, with low resistance, this does not mean that it can be picked up from anywhere as there are plenty of voltage disturbances in the ground plane, which are not related directly to the voltage of the component being measured, which can give measurement results that do not correspond to what the target component is actually affected by. In a good measurement setup, there should not be a large loop area anywhere, and this includes what the probe leads form between + and - points.
I am not going to go in to basics here of what oscilloscope probes to use, or other oscilloscope basics, which I assume anyone playing with measuring power circuits with oscilloscopes should be aware of, but I typically use an isolated probe with 100 MHz bandwidth with short leads, which is more than enough for these kinds of measurements. I have also recently purchased a hand-held 100 MHz scope, and found that to be a good enough tool as well to measure voltages easily on motherboards, as there is no mains power grounding related issues to worry about with any probe.

The output side of the inductors could be measured also, if nothing else is easily accessible, but care has to be taken that the inductor EMI would affect the measurement probe as little as possible. The probe wires can easily pick up elevated levels of disturbance to the measurement due to the inductor's electromagnetic field, depending on how they are positioned in relation to the inductor.

Fascinating! I've learned so much from you, thank you.

I will only add one final comment, before I go:
It does seem that QDI Kinetiz 7E is a much better board overall compared to the Abit KT7A (it is much better designed in circuitry?). However the KT7A does seem to be more common, since some people have 3 of them.

Windows98_guy wrote on 2023-11-20, 10:20:

I will only add one final comment, before I go:
It does seem that QDI Kinetiz 7E is a much better board overall compared to the Abit KT7A (it is much better designed in circuitry?). However the KT7A does seem to be more common, since some people have 3 of them.

One additional important note, regarding QDI Kinetiz 7E, as some people seem to tend to overclock these things for some reason; There is nonexistent support for overclocking in BIOS. Some voltages and bus speeds can be changed, in a very rough and limited fashion. There are some jumper settings on the board for the CPU voltage, bus speed and CPU multiplier, but the multiplier only goes up to 12.5x, and the other settings are also very limited from overclocking standpoint. So, I did the overclocking by leaving the multiplier setting to auto and pinmodded the socket with some very thin wires. Therefore, the Kinetiz 7E is definitely not the board for overclocking. But then again, I think the KT7A is not really a great overclocking board either, and I don't know any KT133A board that is.
At least for me, the main point for any KT133A chipset based board is that it has an ISA connector, and some other desirable features for DOS - Windows 9x retro gaming, and that is about it. The KT133A as a platform starts to be too slow for Windows XP gaming, in my opinion. If the ISA connector is not a requirement, then there definitely exists even better options that enable great gaming experience in DOS, Windows 98, and Windows XP combination running on a single computer.

Finally, the official BIOSes on Kinetiz 7E will only support up to some Thoroughbred Athlon XPs CPUs, up to 2200+ or 2400+, and even that very unofficially... I got the Barton support working by using the rom.by bios patcher and adding the microcodes, and some other patches, like others have done with the Abit KT7A.