There is a thread on the subject here:

VIA C3 Ezra / Ezra-T CPU

An unspecified ASUS slotket is mentioned as an option that works by user [j^aws] but no details on longevity.
The ASUS slotkets that are documented have in common that the lowest voltage that can be jumpered is 1.5v. which is over the limit given for the Ezra but I have no idea if that is damaging.

There were several slotkets that could be jumpered for Cyrix cpu's but my guess is that these typically used auto voltage detection which may or may not have worked flawlessly.

There were also slotkets with the option to select voltages via jumpers starting with +/-1.3 volt.
Some names to look out for are:
Abit SlotKET III
Gigabyte GA-6R7 Pro
Iwill Slocket II
Microstar MS-6905
QDI 370 CPU Card 5.0
Shuttle Adapter Socket 370 HOT-C003
Soltek SL-02A++
Tekram P6TS3

These models do not often pop up on Ebay etc. and it may take some time/effort to score one.

Really appreciate your info, thanks for that. Yeah, in terms of manual configuration of voltage on this Slotket, it seems V1.5 is the lowest. But it also has as "CPU Default" setting, do you know what that means? Would it be risky to try?

CPU default should forward whatever the CPU is asking for to the VRM, so first of all check what that supports. I guess that in theory it could be that the slotket doesn't forward all voltage request lines, in that case asking for a Tualatin voltage (smaller intervals) would result in a slightly higher or lower voltage.

Oh I see. I'm not sure how to confirm that. I'm trying to use a Via C3 Ezra 1.35V on a Gigabyte GA-6BXC rev2.0 motherboard.

static- wrote on 2021-06-20, 18:19:

Oh I see. I'm not sure how to confirm that. I'm trying to use a Via C3 Ezra 1.35V on a Gigabyte GA-6BXC rev2.0 motherboard.

Try to identify the VRM chip (probably near the coils/mosfets near the socket) and look up the datasheet.

Oetker wrote on 2021-06-20, 18:55:

static- wrote on 2021-06-20, 18:19:

Oh I see. I'm not sure how to confirm that. I'm trying to use a Via C3 Ezra 1.35V on a Gigabyte GA-6BXC rev2.0 motherboard.

Try to identify the VRM chip (probably near the coils/mosfets near the socket) and look up the datasheet.

It might be HIP6004CB - 1.3VDC to 3.5VDC
I found a photo here on vogons ( GA-6BXC rev 1.9 ):

Yeah Coppermine compatible VRMs should be able to go down to 1.3v. So as long as you leave the jumpers on the auto setting, all should be well. That the jumpers only go down to 1.5v is probably because that's what Tualatins used. It could also be that by playing with the jumpers other voltages are actually supported. I.e. I wouldn't be surprised if removing JP3 from the 1.5v setting is 1.3v, as this would match the MSI Master slotket.

Thank you so much! I will check my rev2.0 when I get home but I imagine that hasn't changed. Hmm what about the "copper mine" vs "Celeron" jumper? What might that actually do?

Yup it's a HIP6004BCB. From the datasheet looks like 1.3VDC - 3.5VDC as Roman555 said. I guess I'll give this a shot and see what happens!

I have an Nehemiah C3, likely running at either 1.5v or 1.6v, in the same board. Perfectly stable, no issues. To the extent at which it is over volting it I can't say. The cooler isn't hot. I have had it for some time and a few other members are doing the same thing.

static- wrote on 2021-06-20, 21:24:

Thank you so much! I will check my rev2.0 when I get home but I imagine that hasn't changed. Hmm what about the "copper mine" vs "Celeron" jumper? What might that actually do?

Not sure about that jumper. The oldest slotkets only support PPGA Celeron CPUs, so maybe that jumper toggles between that mode and Coppermine pinout. The Coppermine setting is probably correct.

Oetker wrote on 2021-06-20, 20:03:

Yeah Coppermine compatible VRMs should be able to go down to 1.3v. So as long as you leave the jumpers on the auto setting, all should be well. That the jumpers only go down to 1.5v is probably because that's what Tualatins used. It could also be that by playing with the jumpers other voltages are actually supported. I.e. I wouldn't be surprised if removing JP3 from the 1.5v setting is 1.3v, as this would match the MSI Master slotket.

The first wave of Coppermine core Celerons with stepping [B0] ran on 1.5v. Their early 2000 release date corresponds more or less with the start of production of Coppermine slotkets. Tualatins came a year later.

It is well possible that there are a couple of undocumented jumper settings....but the TVC16222 chip on slotkets is a limiting factor and it is well possible (I'm guessing here) that not the same jumper settings as on the MSI are actually made available by ASUS.

I have not been able to find an exact description of what the [Celeron <=> Coppermine] switch actually does.
On the ASUS it consists of just one jumper but on others, like f.e the generic AA370TS, there are two jumpers involved.

My guess is that in some slotket designs the [Celeron <=> Coppermine] switch addresses the AM2, AH4, X4 pin incompatibility between PPGA Celeron and Coppermine.
http://geocities.ws/_lunchbox/pentium3_socket370_mod.html

PARKE wrote on 2021-06-21, 11:38:

The first wave of Coppermine core Celerons with stepping [B0] ran on 1.5v. Their early 2000 release date corresponds more or les […]
Show full quote

Oetker wrote on 2021-06-20, 20:03:

Yeah Coppermine compatible VRMs should be able to go down to 1.3v. So as long as you leave the jumpers on the auto setting, all should be well. That the jumpers only go down to 1.5v is probably because that's what Tualatins used. It could also be that by playing with the jumpers other voltages are actually supported. I.e. I wouldn't be surprised if removing JP3 from the 1.5v setting is 1.3v, as this would match the MSI Master slotket.

The first wave of Coppermine core Celerons with stepping [B0] ran on 1.5v. Their early 2000 release date corresponds more or less with the start of production of Coppermine slotkets. Tualatins came a year later.

It is well possible that there are a couple of undocumented jumper settings....but the TVC16222 chip on slotkets is a limiting factor and it is well possible (I'm guessing here) that not the same jumper settings as on the MSI are actually made available by ASUS.

I have not been able to find an exact description of what the [Celeron <=> Coppermine] switch actually does.
On the ASUS it consists of just one jumper but on others, like f.e the generic AA370TS, there are two jumpers involved.
1110590.jpg
My guess is that in some slotket designs the [Celeron <=> Coppermine] switch addresses the AM2, AH4, X4 pin incompatibility between PPGA Celeron and Coppermine.
http://geocities.ws/_lunchbox/pentium3_socket370_mod.html

Coppermines do run at 1.5v but the VRM 8.4 spec meant for Coppermines prescribes 2.05-1.3v. Of course, there might be motherboards that have a VRM that doesn't go that low.

As I understand it, the TVC16222 has to do with bus or clock related signals and not with the core voltage, i.e. the jumpers shouldn't influence its behavior. The VRM's datasheet will list what request value results in what voltage, and different slotkets will have a different way of jumpering those lines.

Woohoo! It worked!!! I set the slotket to J1-5 to "CPU default" and J6 to "coppermine" and was able to boot the Via C3 Ezra. Actually at first, I had accidentally set this jumper to "Celeron" and it also seemed to work.

One odd thing though, I wasn't able to post at 100MHz. It would show the BIOS but freeze just before the RAM test count. I don't think this is a problem with the slotket though, rather something with the motherboard and CPU? I tried different ram as well but no difference (pc100, pc133).

The label on the C3 CPU shows 100 x 9, so I set the motherboard SW1 to 100MHz and SW2 to 9x.

Fails: 100MHz CPU 66MHz AGP
Works: 66MHz CPU 66MHz AGP

Manual in case anyone is curious:
https://download1.gigabyte.com/Files/Manual/m … nual_6bxc_e.pdf

Thanks for all the help and info, I'm super happy this combination worked. I'll have to do more research and experimentation from here.

*EDIT: I fixed it! Somehow the USB keyboard I had connected was triggering this... after removing it, and using a PS2 keyboard instead, no issues.

I recently bought an older MS6905 Ver:1.1 slotket and it only has manual settings that go down to 1.8V.

Would anyone know if I could use it at 1.35V if I left it on auto? eg. with ga-6bxc + Via C3 Ezra CPU 1.35V?

I have that Asus slotket S370 I mentioned earlier in the thread set to auto and it works this way without issue. I know newer versions of MS6905 support lower voltages, but this older one bottoms out at 1.8V manual it seems. But given what Oetker said about VRM passthrough, perhaps this should work?

I'm not sure if it's coppermine compatible though.

Thanks

I found this, think I could send it in? 🤣
https://www.overclockers.com/msi-replacement-offer/

Presumably since it doesn't' seem to support Coppermine, it can't do voltages this low? 🙁

static- wrote on 2021-07-31, 13:54:

I recently bought an older MS6905 Ver:1.1 slotket and it only has manual settings that go down to 1.8V. *** I'm not sure if it's […]
Show full quote

I recently bought an older MS6905 Ver:1.1 slotket and it only has manual settings that go down to 1.8V.
***
I'm not sure if it's coppermine compatible though.
***
Thanks

Hi.
It's easy to understand if a s370-board (or a slotket) is Cu-Mine-Ready or not: the AM2 pin of S370 shouldn't be connected to GND. You may check it with an ohmmeter.
Also you may download Cu-Mine and Erza datasheets and compare their Voltage Identification pins VID[0:3]. If they are the same - Cu-Mine slotket in "auto voltage mode" just forwards VID signals from CPU into a CPU-core VRM.