Just Friends: 1 - Metropolis: 0, a word of WARNING

Cwejman, Livewire, TipTop Audio, Doepfer etc... Get your euro on!

Moderators: Kent, Joe., luketeaford, lisa

Post Reply
funqpatrol
Common Wiggler
Posts: 176
Joined: Tue Aug 30, 2016 8:14 am
Location: Melbourne

Just Friends: 1 - Metropolis: 0, a word of WARNING

Post by funqpatrol » Sat Oct 29, 2016 2:53 am

Here we go, I've been hesitant to post this but similar stories have emerged and I feel I need to warn ppl now.

What happened: My Just Friends module fried my Metropolis when I tried to modulate the JF's RUN input with it.

Not intending to blame either Whimsical Raps or Intellijel as my knowledge in electronics is limited and it is not clear to me what is the cause of this. But an answer from Intellijel and some discussion on Aussie Wigglers lead to that conclusion: JF's RUN input feeds a -12V jolt into the source when jack is half inserted and the Metropolis looks like is not able to handle that too well. Design flaws?

End result is now I have a dead Metropolis and and a lethal Just Friends that I'm too scared to use.

So be warned ppl that this happened. Please discuss.
Last edited by funqpatrol on Sat Oct 29, 2016 11:02 pm, edited 2 times in total.

User avatar
pieter
Super Deluxe Wiggler
Posts: 1720
Joined: Tue Dec 01, 2015 2:59 pm
Location: UK

Post by pieter » Sat Oct 29, 2016 3:40 am

Oh, wow. I'm sorry for your loss.

Delivering –12V via a half-inserted jack sounds like a fault to me (design or assembly, I can't say). Then again, one would hope that modules would be able to handle the theoretical maximum (minimum) voltage.

Edit: I read the patch the wrong way around. Sending a high voltage into an output (in this case of the Metropolis) is known to be very risky. You can still use JF as long as you remember to patch it first when connecting it to another module, and unpatch JF last. Still, risky...

User avatar
Leverkusen
Super Deluxe Wiggler
Posts: 2348
Joined: Mon Jul 29, 2013 2:26 am
Location: Suhlendorf

Post by Leverkusen » Sat Oct 29, 2016 3:59 am

Oh dear! I just read about this issue on another thread. I am sorry for you and I hope it will solve out somehow - still it's disappointing and a lot of hassle.

I had one of those when they came out but returned it because of software issues and design choices I did not want to deal with but this is really a bummer. I was about to reconsider my decision with the new firmware since I am into monome anyway and 'just type' looks quite interesting, but this holds me back now.

Looking at the big picture my modular experience over the last year has been that quite a few new things did not work as expected due to software issues or hardware design and I had a lot of returns and repairs of new devices.
My personal conclusion was that things might get overly complicated featurewise while the format is evolving more and so the devices are more prone to failure, especially with the QM a cottage industry is able to provide. Maybe the production runs and the range of products are getting to big to handle properly here and there.

So I decided to skip all this complex new stuff, being digital or analogue.

In this special case my electronic knowledge is not that great either but I would assume that it is good practice to not let your modules emit large voltages out of an input AND make them safe with all possible inputs. If I would build modules I would rather make sure that my design is safe in both directions but expect other devices being able to handle my design choices.

Isn't Maths doing it too? being reminded at this frightens me a bit...

User avatar
yghartsyrt
permanent entropy
Posts: 953
Joined: Wed May 04, 2011 1:54 am
Location: Cologne, Germany

Post by yghartsyrt » Sat Oct 29, 2016 4:04 am

Leverkusen wrote:
Isn't Maths doing it too? being reminded at this frightens me a bit...
Maths also puts out a pulse wave on the trigger input. Has never been a problem for me though.

If I remember correctly a lot of function generators do this.

User avatar
Leverkusen
Super Deluxe Wiggler
Posts: 2348
Joined: Mon Jul 29, 2013 2:26 am
Location: Suhlendorf

Post by Leverkusen » Sat Oct 29, 2016 4:15 am

yghartsyrt wrote:
Leverkusen wrote:
Isn't Maths doing it too? being reminded at this frightens me a bit...
Maths also puts out a pulse wave on the trigger input. Has never been a problem for me though.

If I remember correctly a lot of function generators do this.
Neither it has for me. It still make me worry sometimes. I think for my limited understanding the old analogue stuff feels more robust to those features than the shiny new computer modules - even more when it's big clunky through hole technology. :wink:

Dogma
Super Deluxe Wiggler
Posts: 4143
Joined: Wed Aug 28, 2013 10:46 pm

Post by Dogma » Sat Oct 29, 2016 6:44 am

It's a brilliant module - have you contacted whimsical?

I know they just released just type and I'm pretty sure a v2 of the firmware (it's a wav file into a jack al'a mutable)

There support is excellent

User avatar
bemushroomed
Super Deluxe Wiggler
Posts: 2646
Joined: Sat Oct 25, 2014 10:30 am
Location: Sweden

Post by bemushroomed » Sat Oct 29, 2016 7:02 am

Just Fried could be fun name change, if they cant work it out in a firmware fix
:oops:

User avatar
Nofrenchtests
Veteran Wiggler
Posts: 647
Joined: Thu Sep 11, 2014 4:07 pm
Location: Hunter Valley

Post by Nofrenchtests » Sat Oct 29, 2016 7:20 am

Confirmed this one with a multimeter when it was brought up with the Aussie Wigglers, have contacted Gil @ WR about it but haven't had a reply. I'm torn between sending it back and hanging on to it.

User avatar
nearly ghost
Veteran Wiggler
Posts: 645
Joined: Sun Feb 03, 2013 5:23 am
Location: Malta

Post by nearly ghost » Sat Oct 29, 2016 7:47 am

Damn that sucks.. So we need to be careful what is sent into the run input. Anyone know if this can be fixed with a firmware update? How can we know what will be safe and what not?

User avatar
Nofrenchtests
Veteran Wiggler
Posts: 647
Joined: Thu Sep 11, 2014 4:07 pm
Location: Hunter Valley

Post by Nofrenchtests » Sat Oct 29, 2016 7:53 am

nearly ghost wrote:Damn that sucks.. So we need to be careful what is sent into the run input. Anyone know if this can be fixed with a firmware update? How can we know what will be safe and what not?
As far as I can tell, it only happens when the run jack gets shorted/halfway in. If you're not sending it back, always patch it first, always unpatch it last.

User avatar
nearly ghost
Veteran Wiggler
Posts: 645
Joined: Sun Feb 03, 2013 5:23 am
Location: Malta

Post by nearly ghost » Sat Oct 29, 2016 7:58 am

Nofrenchtests wrote:
nearly ghost wrote:Damn that sucks.. So we need to be careful what is sent into the run input. Anyone know if this can be fixed with a firmware update? How can we know what will be safe and what not?
As far as I can tell, it only happens when the run jack gets shorted/halfway in. If you're not sending it back, always patch it first, always unpatch it last.
I'll bear that in mind thanks for the tip

User avatar
aroom
Ultra Wiggler
Posts: 798
Joined: Wed Jun 24, 2015 3:27 am
Location: Switzerland
Contact:

Post by aroom » Sat Oct 29, 2016 8:44 am

I'll just quote Pichenettes from this thread :

viewtopic.php?t=170592
pichenettes wrote:I think the designer(s) of the JF wanted to normalize the "Run" input to -12V so that it was set to a known, easily detectable voltage when nothing is patched in the jack (detecting unpatched jacks is tricky indeed).

The issue of doing it by connecting the "switch" of the jack straight to -12V is that there is a temporary short to -12V during the insertion of the jack, which can cause the module on the other end to send an excess of current, and exceed the current drive capability of an op-amp or GPIO. A more sensible choice would have been to insert a small resistor (2 or 3k is low enough) between the -12V rail and the "switch", so that, in the event of a short during insertion, only (12 - -12) / 2k = 10mV of current is at most sourced or sinked on either sides. In normal operation, this would cause only a small voltage drop (100/102 x -12 = -11.7V).

I made a similar mistake early in the genesis of Tides, though I found it before the module went in production.

Meanwhile, the problem can be avoided by first patching the jack on the JF side, and only then patching the other side.

User avatar
acgenerator
Ultra Wiggler
Posts: 784
Joined: Tue Jun 30, 2015 8:44 am
Location: USA

Post by acgenerator » Sat Oct 29, 2016 8:56 am

Thank you for the warning. It might be better to change the title so folks can see it's an issue though.

I can confirm the -12v on partial insert using Division 6's Voltmeter (why is this module not more popular?) My manual / scroll indicated my module was part of the original batch.

meanwhile I'm going to cover up that socket until an answer is found. Not sure if a firmware patch is the answer as the circuit flow is likely physically disrupted.

s simple as a diode to restrict the flow might work but this might put limitations on the signals that could be sent.

A physical switch to switch to Run mode after the patch is in place might be an option but it would be a board redesign.

User avatar
thisisprisma
Wiggling with Experience
Posts: 274
Joined: Thu Dec 20, 2012 5:16 am
Location: Spain
Contact:

Post by thisisprisma » Sat Oct 29, 2016 9:48 am

sorry to hear about your loss and thanks for letting us know, so we will have to be careful and always patch into just friend first and disconnect last from RUN.

User avatar
acgenerator
Ultra Wiggler
Posts: 784
Joined: Tue Jun 30, 2015 8:44 am
Location: USA

Post by acgenerator » Sat Oct 29, 2016 9:49 am

As a side note to manufacturers:
- This is why we need to have a stricter standards around Eurorack.
- Your testing should include unexpected inputs and design should include safeguards to minimize damage. You can't prevent everything but a basic set of tests should be part of the standard. This should include:
a) any jack being able to withstand the full +12/-12 volt range (regardless of function) coming from both CV and Audio ranges.
b) multiple supply types, reversed cable, bad grounding, power starvation, noisy power supplies.
- If your module has unusual voltage/power drops or spikes that should be called out in the manuals (e.g. Vacuum Tubes draw at start-up). This of course requires you write a manual.
- Same thing goes for calling out any restrictions where your module is incompatible with anything that is falls under the "standards".

As a side note to users:
- Get in the habit of setting up a testing routine for new modules. It should be part of safeguarding your investments. This especially applies to DIY, Betas, new releases, and used modules. Test thoroughly with your replaceable stuff first! We'll all experience a broken module sooner or later - at least hedge your bets.
- Part of the responsibility falls in our court to understand that no matter the amount of testing and design check, inevitably something will go wrong in the field eventually. Manufacturers won't be able to test for all future scenarios as dozens if not hundreds of new modules will get released each year. The manufacturers can't possibly test with every module out there. Being on the "bleeding edge" requires a bit of accepted risk.

User avatar
joem
Ultra Wiggler
Posts: 965
Joined: Thu Mar 05, 2015 11:07 pm
Location: NYC

Post by joem » Sat Oct 29, 2016 1:37 pm

acgenerator wrote:As a side note to manufacturers:
- This is why we need to have a stricter standards around Eurorack.
- Your testing should include unexpected inputs and design should include safeguards to minimize damage. You can't prevent everything but a basic set of tests should be part of the standard. This should include:
a) any jack being able to withstand the full +12/-12 volt range (regardless of function) coming from both CV and Audio ranges.
b) multiple supply types, reversed cable, bad grounding, power starvation, noisy power supplies.
- If your module has unusual voltage/power drops or spikes that should be called out in the manuals (e.g. Vacuum Tubes draw at start-up). This of course requires you write a manual.
- Same thing goes for calling out any restrictions where your module is incompatible with anything that is falls under the "standards".
Unless you're also proposing a standards committee that certifies modules, I'm not really sure this will help. Self certification is just as much of a gamble as things are now.

And if you are actually proposing a standards committee and certification process, then, well, I strongly do not want that bureaucracy in eurorack.

User avatar
acgenerator
Ultra Wiggler
Posts: 784
Joined: Tue Jun 30, 2015 8:44 am
Location: USA

Post by acgenerator » Sat Oct 29, 2016 3:20 pm

joem wrote:
acgenerator wrote:As a side note to manufacturers:
- This is why we need to have a stricter standards around Eurorack.
- Your testing should include unexpected inputs and design should include safeguards to minimize damage. You can't prevent everything but a basic set of tests should be part of the standard. This should include:
a) any jack being able to withstand the full +12/-12 volt range (regardless of function) coming from both CV and Audio ranges.
b) multiple supply types, reversed cable, bad grounding, power starvation, noisy power supplies.
- If your module has unusual voltage/power drops or spikes that should be called out in the manuals (e.g. Vacuum Tubes draw at start-up). This of course requires you write a manual.
- Same thing goes for calling out any restrictions where your module is incompatible with anything that is falls under the "standards".
Unless you're also proposing a standards committee that certifies modules, I'm not really sure this will help. Self certification is just as much of a gamble as things are now.

And if you are actually proposing a standards committee and certification process, then, well, I strongly do not want that bureaucracy in eurorack.
Not so much a certification process but an updated standard where folks can voluntarily agree to adhere to. Most industries have that whether it is ISO or not. In fact, most electronic instruments already have some standards they try to meet anyways -- general electricity standards, MIDI, cable formats, the 19" rack, etc.

There are certainly enough times a good number of folks are in the same room that a conversations can start to be had... NAMM, Knobcon, Superbooth, etc. Folks that opt in can self-certify that they meet that standard (or at least what version of it). Hell we don't even need to call it a "standard" if that chafes people the wrong way - call it "industry best practices", "guidelines", "Euro-crack club rules" or whatever.

The consequence of not agreeing to play well together will ultimately lead to the market forces deciding for them.

User avatar
Leverkusen
Super Deluxe Wiggler
Posts: 2348
Joined: Mon Jul 29, 2013 2:26 am
Location: Suhlendorf

Post by Leverkusen » Sat Oct 29, 2016 3:41 pm

acgenerator wrote:
There are certainly enough times a good number of folks are in the same room that a conversations can start to be had... NAMM, Knobcon, Superbooth, etc. Folks that opt in can self-certify that they meet that standard (or at least what version of it). Hell we don't even need to call it a "standard" if that chafes people the wrong way - call it "industry best practices", "guidelines", "Euro-crack club rules" or whatever.

The consequence of not agreeing to play well together will ultimately lead to the market forces deciding for them.
As far as I know some of this was tried but did not work out pretty good - just don't start with how control voltage ranges should be.

But anyway, what you are describing is more that devices should be designed proper and safe, which is not a standard.

Plus I think you overestimate the power and wisdom of the so called market forces in making sensitive decisions.

funqpatrol
Common Wiggler
Posts: 176
Joined: Tue Aug 30, 2016 8:14 am
Location: Melbourne

Post by funqpatrol » Sat Oct 29, 2016 4:37 pm

pichenettes wrote:I think the designer(s) of the JF wanted to normalize the "Run" input to -12V so that it was set to a known, easily detectable voltage when nothing is patched in the jack (detecting unpatched jacks is tricky indeed).

The issue of doing it by connecting the "switch" of the jack straight to -12V is that there is a temporary short to -12V during the insertion of the jack, which can cause the module on the other end to send an excess of current, and exceed the current drive capability of an op-amp or GPIO. A more sensible choice would have been to insert a small resistor (2 or 3k is low enough) between the -12V rail and the "switch", so that, in the event of a short during insertion, only (12 - -12) / 2k = 10mV of current is at most sourced or sinked on either sides. In normal operation, this would cause only a small voltage drop (100/102 x -12 = -11.7V).

I made a similar mistake early in the genesis of Tides, though I found it before the module went in production.

Meanwhile, the problem can be avoided by first patching the jack on the JF side, and only then patching the other side.

Ok so there is a short, great. I should mention that when it happened the wogglebug's blue tempo led went very bright and steady as well. Lucky it didn't damage other modules. I'd really like to ear what whimsical has to say now.

Also from a Swedish site, another dude had the same exact unfortunate experience:

http://www.99musik.se/showthread.php?33 ... ht=friends

funqpatrol
Common Wiggler
Posts: 176
Joined: Tue Aug 30, 2016 8:14 am
Location: Melbourne

Post by funqpatrol » Sat Oct 29, 2016 4:55 pm

bemushroomed wrote:Just Fried could be fun name change, if they cant work it out in a firmware fix
:oops:
aha More seriously could it be fixed with a firmware update though? sounds more like a hardware issue.
Last edited by funqpatrol on Sat Oct 29, 2016 6:24 pm, edited 2 times in total.

User avatar
PM33AUD
Destroyer of BSP
Posts: 982
Joined: Sat Apr 23, 2011 7:28 pm
Location: USA aka dumpster fire

Post by PM33AUD » Sat Oct 29, 2016 5:01 pm

Standards are only as good as those who follow them and more importantly, those who have created them. That is far more difficult than it may seem, especially when you are dealing with humans with varying levels of ego.

The fortunate thing is that good engineering can get around most of the lack of standards. It *really* sucks to have to design this way, but my response is: 'it is what it is'... and 'everyone, myself included, likes the wild wild west.'

Regarding this short if it is indeed a direct short on the normalisation, yes, just add a resistor. Should be an easy fix/mod, even on a SMT board and then the next products can just do it this way going forward.

As a side note, most device outputs that are 'short circuit protected' cannot deal with shorts to something other than ground all that well (most chip manufacturers don't even do tests or provide ratings for this). Worst case you have +12 driving into -12, which is a lotta potential and lots o current that can have a go. I like using output stages that have built-in thermal protection which is pretty commonplace on most OAs out these days. MCUs/digital IO always need appropriate clamping/current limiting.

User avatar
Dr. Sketch-n-Etch
Super Deluxe Wiggler
Posts: 7952
Joined: Thu Jan 21, 2010 1:46 pm
Location: Vancouver, BC, Canada

Post by Dr. Sketch-n-Etch » Sat Oct 29, 2016 5:24 pm

I don't really understand why a -12V pulse would damage anything on any module. It is more likely that it sends a blast of high current which can damage something. Olivier is probably right, that the jack switch is connected directly to the -12V rail, in which case the current it can send is more or less whatever the power supply can provide. Normalling it to ground would have been a saner choice. Severing the normalled connection might be the best option to avoid any future issues.

However, (I didn't design it and I don't have the schematics on my computer, but) I have to assume that the Metropolis's output is buffered and current limited with output resistors. I don't really understand how it can be damaged in this way.
Composting the drones will ensure the survival of the elite.

User avatar
joem
Ultra Wiggler
Posts: 965
Joined: Thu Mar 05, 2015 11:07 pm
Location: NYC

Post by joem » Sat Oct 29, 2016 5:26 pm

[actually nevermind... deleted post]

User avatar
wsy
Super Deluxe Wiggler
Posts: 2719
Joined: Wed Aug 24, 2011 10:20 pm
Location: near Boston Massachusetts USA

Post by wsy » Sat Oct 29, 2016 6:49 pm

yghartsyrt wrote:
Leverkusen wrote:
Isn't Maths doing it too? being reminded at this frightens me a bit...
Maths also puts out a pulse wave on the trigger input. Has never been a problem for me though.

If I remember correctly a lot of function generators do this.
Not quite the "trigger" input - it's on the "integrate in" input, and only when you have "CYCLE" turned on.
Hasn't caused a problem yet, but yeah, it kinda gives one pause and a bit of concern.

It's fairly potent- around 10 volts, and seems to have quite a bit of current sourcing (i.e. it's not just through a 100K resistor,
knocking down the current to a 0.01 mA or so. It's more than enough to override what a Rene can put out on it's clock outputs.

Good to know.... and good justification to design modules so that inputs AND OUTPUTS can sustain continuous
short to ground, V++, and V--. There were some unbelievers here, IIRC....

- Bill
"Life is short. But we can always buy longer patch cords" - Savage

tbecker
Veteran Wiggler
Posts: 640
Joined: Thu Aug 21, 2014 9:48 am

Post by tbecker » Sat Oct 29, 2016 7:19 pm

[quote="PM33AUD"]Standards are only as good as those who follow them and more importantly, those who have created them. That is far more difficult than it may seem, especially when you are dealing with humans with varying levels of ego.

The fortunate thing is that good engineering can get around most of the lack of standards. It *really* sucks to have to design this way, but my response is: 'it is what it is'... and 'everyone, myself included, likes the wild wild west.'

Regarding this short if it is indeed a direct short on the normalisation, yes, just add a resistor. Should be an easy fix/mod, even on a SMT board and then the next products can just do it this way going forward.

As a side note, most device outputs that are 'short circuit protected' cannot deal with shorts to something other than ground all that well (most chip manufacturers don't even do tests or provide ratings for this). Worst case you have +12 driving into -12, which is a lotta potential and lots o current that can have a go. I like using output stages that have built-in thermal protection which is pretty commonplace on most OAs out these days. MCUs/digital IO always need appropriate clamping/current limiting.[/quote]

I have reading this thread closely, but for me it sounds like the run jack should be avoided or pre patched. I think whimsical could help to detail a fix and I am sure they will soon. FYI, even Cwejman modules like the MMF-6 have had issues that required a field fix. In this case, it sounds like a simple in field soldered resistor could fix things which I can do if Gil wants to point us to the right terminals. Based on my interactions with whimsical, I think this service will be provided soon once the 'fix' tests correct for others that don't like soldering irons.

Post Reply

Return to “Eurorack Modules”