Page 1 of 7

RetroStone boot problem identified

Posted: Wed 10 Oct 2018 16:16
by Admin
Hey guys,
I just found the reason for the specific boot problem all retrostone are having sometimes.

It usually is as follow :
- 10-20% of time the unit will stay on blue/white LCD forever
- 90-80% of time the unit boot OK.

The good news is that I found the issue and that future retrostones will be corrected. The bad news is that it's a hardware issue and to fix it you need to remove a resistor. (same as audio filter fix).
The resistor to remove is R23 just under the SD card reader. It should be removed and replaced with solder. Or you can just drop a ball of solder on top of the resistor, no need to remove the resistor. Or as someone suggest after you can just use conductive glue : put a blob of glue on the resistor and let it dry.

I will post a video asap.
Very sorry about this additional imperfection :(

Re: RetroStone boot problem identified

Posted: Wed 10 Oct 2018 17:13
by RetroJack
How did you discover the fix?

Re: RetroStone boot problem identified

Posted: Wed 10 Oct 2018 18:48
by Daniel
I am very disappointed with the Retrostone where I was so excited by the Raspiboy. First the problem with the Dpad sensitivity, secondly the problem with the much too soft sound from the headphones and now the boot problem. I do not have a soldering iron at home and I do not have the skills to fix these problems by soldering and I think most users will think so.
I think it's extremely unfair to leave the troubleshooting to users.

Re: RetroStone boot problem identified

Posted: Wed 10 Oct 2018 19:04
by Bart0
Is it going to solve the boot problem for sure? Could you post a pic or a video? I would try. But I wouldn't want to risk again if it's not for 100% the source of this boot issues.

Re: RetroStone boot problem identified

Posted: Wed 10 Oct 2018 19:44
by dr_bunsen
@Pierre How much experience have you had with the unit after the fix? I'm concerned that this may introduce other issues or not be a proper fix. What's the science behind it, why does it fix the booting?

Re: RetroStone boot problem identified

Posted: Wed 10 Oct 2018 21:39
by bedfellow
Can you assure those who attempt to perform this resistor modification that you will honour warranties if it goes wrong?

I feel that the product is lacking quality control, with the D-pad issue and now this resistor issue.

Re: RetroStone boot problem identified

Posted: Thu 11 Oct 2018 10:58
by Admin
I understand your disapointments guys... I'm really really sorry about this mistake. I know there are several things that were not correct/perfect with the first run of Retrostone and that's really embarassing for me.

The thing is that for the RetroStone the PCB is a whole motherboard, including CPU, RAM and all peripherals. There are just so much components and schematic is so complex that I have not found these problems sooner :/

I would love to tell you that I could send everyone replacement or fix them all, but the amount of money calling back all the retrostones, rework them and reship them is not an order of magnitude the project can handle... I wish I was selling tens of millions of unit with 80% of margins as some big companies out there, but that's just not the case, I'm a very small business with low volume and margins.


About this fix, this is for the specific boot problem as follow :
- 10-20% of time the unit will stay on blue/white LCD forever (if you have a HDMI connected you should see the message as on this picture "MMC: no card present" ...)
- 90-80% of time the unit boot OK.

The % seems to depend on units, some it's 5% for some others it's closer to 25 or 30% of bad boot.

This is not confirmed to work for units that crashes/freeze/boot errors with text/black LCD (I'm testing it today to see if that also fix these). If you have such unit have a look here : viewtopic.php?f=19&t=1760

About why I found it out, is that this resistor has no reason to be there. It's on the SD card detect trace. There is a pull up resistor on SD-detect to 3.3V, but there was a resistor connected to ground too. So instead of SD-detect being pulled between 3.3v and ground it is pulled between 3.3V and 1.7V. Now that I found it it seems pretty obvious but before finding it it was not so obvious... Reason why I put this 10k resistor in the first place is because it was in the reference schematic, but the pull up resistor value was a special component with multiple resistor of higher value(100k), so on reference schematic it was pulled between 3.3V and 0.3V (so almost ground). As other H3 boards did not used this special component I did the same as other H3 boards and replaced it with a 10k pull up resistor. Where the mistake is because other H3 boards do not have the 10k resistor to ground...

I tested to boot about 50 times on my test unit without a single boot fail. Today I will try to do the fix on crash/freeze/boot-errors-with-text/boot-error-black-LCD RetroStones and I will check if that fix the issues as well.
Doing so I will shoot a video and upload it online.

PS: if you damage the unit while doing so the warranty will be extended of course.

Re: RetroStone boot problem identified

Posted: Thu 11 Oct 2018 13:13
by Brandon
That is great news Pierre,

Thanks for all the updates! I will remove and bridge the resistor later today and let everyone know if it helps.

Re: RetroStone boot problem identified

Posted: Thu 11 Oct 2018 15:34
by Admin
Here's the video : https://www.youtube.com/watch?v=I-dIhERPuZw

Also I tested on crash/freeze/boot-errors-with-text/boot-error-black-LCD RetroStones and it seems it does not solve the issue. If you have such a unit it should be replaced. (viewtopic.php?f=19&t=1760)

Re: RetroStone boot problem identified

Posted: Thu 11 Oct 2018 16:21
by ruffnutts
Thanks for the video and fix, could we use conductive PCB glue to stort the risistor for people that can't solder, I'm almost sure it would work, Just a blob on top 8-)