SDRAM board

https://github.com/MiSTer-devel/Main_MiSTer/wiki

Moderators: Mug UK, Zorro 2, spiny, Greenious, Sorgelig, Moderator Team

Locked
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

Sorgelig wrote:
digitron wrote:Do you have a link to the 2.5 order on pcbway? The 2.4 boards are $50.
PCBWay has no special cost for smaller board. So both 2.4 and 2.5 cost the same. That's why there is no v2.5 on PCBWay.
Hi Sorgelig, thank you for this information, think I found it from the github: https://jlcpcb.com/quote#/?orderType=1& ... aseNumber=
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

Does the voltage matter on the 1uF caps? eg. 10v, 16v, 50v? I don't see it in the schematic, sorry.
zerohimself
Retro freak
Retro freak
Posts: 16
Joined: Wed Dec 12, 2018 10:09 pm

Re: SDRAM board

Post by zerohimself »

digitron wrote:Does the voltage matter on the 1uF caps? eg. 10v, 16v, 50v? I don't see it in the schematic, sorry.
Generally as long as the voltage is above double the maximum that can ever be supplied for the circuit, you are in good shape (or so I have been told, If I am mistaken, please correct me!). Meaning that you should be able to use any of the above ratings. If I were you I would go for whichever is the cheapest.
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

zerohimself wrote:
digitron wrote:Does the voltage matter on the 1uF caps? eg. 10v, 16v, 50v? I don't see it in the schematic, sorry.
Generally as long as the voltage is above double the maximum that can ever be supplied for the circuit, you are in good shape (or so I have been told, If I am mistaken, please correct me!). Meaning that you should be able to use any of the above ratings. If I were you I would go for whichever is the cheapest.
Thanks!

https://www.mouser.com/ProductDetail/KE ... 3YPCDvc%3D
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

digitron wrote:Will this IC Inverter work/fit?

https://lcsc.com/product-detail/74-Seri ... C8207.html
it seems yes.
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

Sorgelig wrote:
digitron wrote:Will this IC Inverter work/fit?

https://lcsc.com/product-detail/74-Seri ... C8207.html
it seems yes.
Thanks Sorgelig <3
jft
Atari nerd
Atari nerd
Posts: 49
Joined: Sun Oct 28, 2018 12:07 pm

Re: SDRAM board

Post by jft »

digitron wrote:Will this IC Inverter work/fit?

https://lcsc.com/product-detail/74-Seri ... C8207.html
I used that part on my boards. It works.
Mlaurito
Atari User
Atari User
Posts: 30
Joined: Mon Jul 22, 2019 9:07 pm

Re: SDRAM board

Post by Mlaurito »

Sorgelig wrote:try all 10uf caps on 64mb module.
Hello! Do you think this is still the best suggestion for 64mb module on a 1.1xs board ? I have all the parts already and a spare unused board so was thinking of assembling one, despite no being 100% supported, thanks!
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

It's not a good idea to make new 64MB modules. Go 128MB instead.
digitron
Atari maniac
Atari maniac
Posts: 92
Joined: Sat Apr 20, 2019 7:27 pm

Re: SDRAM board

Post by digitron »

Whats the latest/greatest capacitor configuration? Will schematic be updated? ty
drj3rk
Atari freak
Atari freak
Posts: 62
Joined: Tue May 14, 2019 10:12 pm

Re: SDRAM board

Post by drj3rk »

zerohimself wrote:
digitron wrote:Does the voltage matter on the 1uF caps? eg. 10v, 16v, 50v? I don't see it in the schematic, sorry.
Generally as long as the voltage is above double the maximum that can ever be supplied for the circuit, you are in good shape (or so I have been told, If I am mistaken, please correct me!). Meaning that you should be able to use any of the above ratings. If I were you I would go for whichever is the cheapest.
It depends on the purpose of the capacitor, but that's a good, safe way to approach it. It's good to have a bit of room over the typical supplied voltage, just in case. I wouldn't say you need over double, unless you're working with AC signals, and counting both directions (from 0V). In the case of a single-supply bypass capacitor, (for example 0V to +V) just something over +V will work fine.

Another thing to note, since these boards are all likely using MLCC type capacitors, it's probably a good idea to get C0G/NP0 rated, as the lower rated MLCCs can oscillate in some conditions. There are a couple of papers out there on this. I can't remember where I read them off the top of my head though.
TWINKYTWONKA
Atariator
Atariator
Posts: 25
Joined: Wed Dec 12, 2018 8:18 pm

Re: SDRAM board

Post by TWINKYTWONKA »

I'd just like to add something that may help those getting low test scores on the 128mb sdram boards. The part that can really make a difference to clock speed is the quality of the soldering on the caps to the pcb. Make sure you take your time and do a really professional job and you'll get 150 mhz pretty much every time. The ram chips at least in my experience are no different to the soldering of the 32mb chips.
Cubanator
Atarian
Atarian
Posts: 2
Joined: Fri Oct 04, 2019 4:36 pm

Re: SDRAM board

Post by Cubanator »

witchmaster wrote:Got the parts and soldered a 128MB SDRAM card. I used Sorgeligs capacitor configuration (shown below) and it seems to work fine (stable @150MHz).
top: 10uf, 0.1uf, 10uf, 0.1uf
bottom: 10uf, 1uf, 1uf, 10uf, 10uf, 1uf, 1uf, 10uf
@witchmaster
Could you help clarify this? Which one is the top and the bottom?
Thank you!
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

Cubanator wrote:
witchmaster wrote:Got the parts and soldered a 128MB SDRAM card. I used Sorgeligs capacitor configuration (shown below) and it seems to work fine (stable @150MHz).
top: 10uf, 0.1uf, 10uf, 0.1uf
bottom: 10uf, 1uf, 1uf, 10uf, 10uf, 1uf, 1uf, 10uf
@witchmaster
Could you help clarify this? Which one is the top and the bottom?
Thank you!
how about read what you've quoted?
User avatar
witchmaster
Captain Atari
Captain Atari
Posts: 150
Joined: Wed Jun 15, 2011 10:50 am

Re: SDRAM board

Post by witchmaster »

Cubanator wrote:
witchmaster wrote:Got the parts and soldered a 128MB SDRAM card. I used Sorgeligs capacitor configuration (shown below) and it seems to work fine (stable @150MHz).
top: 10uf, 0.1uf, 10uf, 0.1uf
bottom: 10uf, 1uf, 1uf, 10uf, 10uf, 1uf, 1uf, 10uf
@witchmaster
Could you help clarify this? Which one is the top and the bottom?
Thank you!
Use the method of elimination, one edge has 4 caps, the other has 8.
Cubanator
Atarian
Atarian
Posts: 2
Joined: Fri Oct 04, 2019 4:36 pm

Re: SDRAM board

Post by Cubanator »

Sorgelig wrote:
Cubanator wrote:
witchmaster wrote:Got the parts and soldered a 128MB SDRAM card. I used Sorgeligs capacitor configuration (shown below) and it seems to work fine (stable @150MHz).
@witchmaster
Could you help clarify this? Which one is the top and the bottom?
Thank you!
how about read what you've quoted?
Yes, you and @witchmaster are correct, my apologies for missreading/missinterpreting. I am new to all this (electronics) and my brain farted!
Thank you!
Shaneus
Atariator
Atariator
Posts: 26
Joined: Fri Mar 15, 2019 12:07 am

Re: SDRAM board

Post by Shaneus »

Can anyone tell me what the difference is between the XSD or the XXS boards? Any speed differences or is it just re: mounting?

Best I can tell, it looks like the XSD is more suitable with existing case designs but the XXS is not (but XXS is better if you're not using a standard case).
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

Shaneus wrote:but the XXS is not (but XXS is better if you're not using a standard case).
No, it's not better. It's just more convenient for seller to solder. There is no other reason for XXS. Name XXS is arguable as it protrudes to the side making the overall size of MiSTer larger.
antoniovillena
Atari User
Atari User
Posts: 43
Joined: Mon Oct 24, 2016 9:11 pm

Re: SDRAM board

Post by antoniovillena »

Sorgelig wrote:
Shaneus wrote:but the XXS is not (but XXS is better if you're not using a standard case).
No, it's not better. It's just more convenient for seller to solder. There is no other reason for XXS. Name XXS is arguable as it protrudes to the side making the overall size of MiSTer larger.
There is other reason. To fit my custom cases. With the XSD ones the case must be double height.
You do not have the required permissions to view the files attached to this post.
zerohimself
Retro freak
Retro freak
Posts: 16
Joined: Wed Dec 12, 2018 10:09 pm

Re: SDRAM board

Post by zerohimself »

Ok, I have made a LOT of the 128mb modules at this point. My first order of PCB's consisted of 80 each of 4 layer red and blue XSD 2.4 pcb's from JLCPCB, with JLC7628 Impedance, on FR4-Standard Tg 130-140C with HASL (lead). Out of those PCB's I was getting a lot of failures. like on the blue PCB's, 4/8 were failing on average, and on the reds 2/8 were failing on average. Solder temp was 340C, using kester 275 solder and kester 959t no clean flux. I did a run of -6tins and -7tins. both with the same kind of failure rates.

(to clairify, i consider a fail to have ANY errors at 140 or lower, or to have issues with archie, memtest, minimig, or neogeo core running bad apple)

I talked about it, and everyone kept saying it's the solder job, etc. I just set my failures aside and kept going to meet orders. I Just recently ordered a batch of the XSD 2.5's from another vendor, with essentially the same settings (They had no option for impedance). So far I only have 1/24 failed modules (12 7tins, and 12 6tins). I have no idea what could be different. All of my materials and processes have stayed the same. I have no idea what to suspect or examine to figure out what is going on here, But i really felt it necessary to speak up in case anyone else is having some significant issues like this.

So far, the chips i have recycled appear to be passing perfectly on the new 2.5 PCB's as well. I would have posted sooner, but I really wanted to validate my results before I started crying with joy that i found my problem.

Does anyone have any ideas of what could be going on with the original batches of PCB's? I'm really curious to know why they were failing at such an incredible rate.
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

What you mean by "fail". Doesn't such module work at all, or doesn't pass some frequency?
v2.4 has smaller clearance between solder pads and ground plane. v2.5 has larger clearance making it easier to solder. Not sure if this can be a reason.
zerohimself
Retro freak
Retro freak
Posts: 16
Joined: Wed Dec 12, 2018 10:09 pm

Re: SDRAM board

Post by zerohimself »

Sorgelig wrote:What you mean by "fail". Doesn't such module work at all, or doesn't pass some frequency?
v2.4 has smaller clearance between solder pads and ground plane. v2.5 has larger clearance making it easier to solder. Not sure if this can be a reason.
My process for testing as a seller goes like this:
1. memtest at 150mhz (pass if it has some errors)
2. memtest at 140mhz (fail if it has any errors)
3. test boot with archie core (fail if it hangs)
4. test boot and play a game with minimig (fail if has hanging issues with known good games)
5. test boot and watch for visual distortions with the neogeo core and bad apple rom (fail if it has any video compression errors)

Most of my fails come from the memtest, and on those (2.4) pcb's, they fail to pass any memtest above 120 (however they do work with archie,minimig,neogeo, and all other tested cores).
So far with the XSD 2.5, I have only had one problem out of around 30 modules and 90% seem to pass at 150mhz as well as 140mhz.

It could possibly be a soldering issue(but i have made a LOT of sdram modules at this point, most without issues), but i really suspect a fault with the pcb at time of manufacture. I'm curious if anyone has an explaination of what JLCPCB (or me) did wrong that is causing this..

Do you recommend any changes to my testing process?
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: SDRAM board

Post by Sorgelig »

I recommend to try to desolder the chips from faulty module, then inspect under magnifying glass the soldering of connector. Then clean it again from flux (by IPA or Ethanol). Then solder back the chips again. Inspect by magnifying glass if there any hanging legs of chips. If any soldering point of connector has bubble, then RAN chip won't lay flat, and legs may hang on the air. It's not a problem actually, but soldering of chip must be careful as some legs won't be soldered. Although you've told the cores work fine (besides the memtest), it's better to try to resolder just to make sure everything is right.
I will check with one seller who soldered many 2.4 and 2.5 if he had noticeable failure rates..
JimDrew
Atari Super Hero
Atari Super Hero
Posts: 865
Joined: Mon Nov 04, 2013 5:23 pm

Re: SDRAM board

Post by JimDrew »

I have had similar issues with 4 layer SDRAM boards from JLCPCB. I think the issue lies in the layer stack height. I have not cross-sectioned a board to see the layer spacing, but it is possible that the layers are not evenly spaced and this affects the impedance and cross-talk. I know that some PCB houses uses a much larger center FR4 layer and thinner outer FR4 layers. Your design will dictate what type of board you need.

One other thing that I forgot to mention - you have to specify the layer order when ordering 4+ layer boards through JLCPCB. If you don't, the inner layers can be swapped. That could be a problem if you are trying to isolate high frequency signals between either outer layers.
I am the flux ninja
Locked

Return to “MiSTer”