VOGONS


First post, by tauro

User metadata
Rank Member
Rank
Member

So I finally got this cache module:

The attachment cache.jpg is no longer available

According to what I've read this module (01S3) is for 3.3v CPUs only.

Does that mean that the cache modules get the same voltage as the CPU? Sometimes I do some testing and set my 3.3v CPUs to 4v... will the cache chips also get 4v? if so, will they withstand it?

Since they're not easy to come by, I want to be careful with this cache module.

Here's some info about the chips:

TC55V328AJ-15 - Maximum rating for VDD is 4.6v. Recommended max 3.6v.
KM68V257J-15 - Both C and E have a maximum rating voltage on any pin of 4.6v and recommended max voltage is 3.6v. I don't know about my "J" chips, since I couldn't find the precise datasheet.

What's your advice? Would it be safe to use this cache module with a 4v CPU?

Reply 1 of 5, by paradigital

User metadata
Rank Oldbie
Rank
Oldbie

Well, the cache slot will be provided with power regardless of a module being installed or not. So it should be pretty simple to check if the voltage provided to the cache is directly tied to the CPU core voltage or not.

The module certainly has no voltage regulation on it.

Reply 2 of 5, by majestyk

User metadata
Rank Oldbie
Rank
Oldbie

The two transistors of the VRM circuit are working parallel for the same CPU core-voltage. This voltage is also feeding the "COAST slot".
If you set the VRM to 4V, the cache chips will operate with 4V also.

Reply 3 of 5, by tauro

User metadata
Rank Member
Rank
Member
majestyk wrote on 2023-09-09, 15:38:

The two transistors of the VRM circuit are working parallel for the same CPU core-voltage. This voltage is also feeding the "COAST slot".
If you set the VRM to 4V, the cache chips will operate with 4V also.

This board has only one transistor and an empty space... I suppose that means that it can provide less current than the one with two.

paradigital wrote on 2023-09-09, 14:43:

Well, the cache slot will be provided with power regardless of a module being installed or not. So it should be pretty simple to check if the voltage provided to the cache is directly tied to the CPU core voltage or not.

I measured it and it's getting the same voltage as the CPU.

The question that remains to be answered is the one in the topic title.
I'm sure that 5v would fry the chips, but is it safe to use this cache module at 4v or better not to?
I haven't found the datasheet for KM68V257J-15 but probably somebody knows.

Reply 4 of 5, by majestyk

User metadata
Rank Oldbie
Rank
Oldbie

The Toshiba chip is rated for 3.0 - 3.6V, so overvolting it by 20% would mean operation out of specs.
I guess it´s similar for the Samsung chips.

If you can replace any defective chips yourself - just run it with 4V, if you want to keep this rare cache stick original, try to run your cpu stable at 3.6V with better cooling for example. Is your cpu an overvolted 3.3V one or a Cyrix?

Reply 5 of 5, by tauro

User metadata
Rank Member
Rank
Member
majestyk wrote on 2023-09-09, 16:24:

The Toshiba chip is rated for 3.0 - 3.6V, so overvolting it by 20% would mean operation out of specs.
I guess it´s similar for the Samsung chips.

I'm under the same impression. It would be out of specs but they should tolerate 4v, I guess.

majestyk wrote on 2023-09-09, 16:24:

If you can replace any defective chips yourself - just run it with 4V, if you want to keep this rare cache stick original, try to run your cpu stable at 3.6V with better cooling for example. Is your cpu an overvolted 3.3V one or a Cyrix?

I'm currently doing some experiments with a couple of Am5x86 that run stable @180MHz, 4v (Quake timedemo OK). S370 heatsink + fan.

Finally I ran some tests with 4v and the cache chips are fine with it, but I don't think that I'm going to use it at 4v long term.