Broken Midiverb II: Troubleshooting Tips?

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: luketeaford, Joe., lisa, Kent

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Fri Jul 31, 2020 11:11 pm

id probably start with the data lines to the DAC or the SRAM. first just check that they all are toggling. you can then see what values are being sent. you probably dont have enough probes for a full 16b data word, but start with the top bits and see what they say.
openmusiclabs.com

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 1:37 am

Thanks, I checked and all the DAC data lines and SRAM lines are toggling.

I also checked the clock on a few of the logic chips and it looked correct, as far as I could tell...

I'm not sure what to look for in the DAC data exactly, but this is the first 20 samples of bits 1-8, sampled at 100kHz
1,0,0,0,0,0,0,0
1,0,0,0,1,1,1,1
0,1,1,1,0,1,1,1
0,1,1,1,1,1,1,1
0,1,1,1,1,1,1,0
0,1,1,0,1,1,1,1
0,1,1,1,1,1,1,1
0,1,1,1,1,1,1,1
1,0,0,1,1,1,1,1
1,0,0,0,0,0,0,0
1,0,0,0,0,0,0,0
0,1,1,1,1,1,1,1
0,1,1,1,1,1,0,1
0,1,1,1,1,1,1,1
1,0,0,0,0,0,1,1
0,1,1,1,1,0,1,1
0,1,1,1,1,1,1,1
1,0,0,0,1,1,1,1
1,0,0,0,0,1,1,1

User avatar
EATyourGUITAR
has no life
Posts: 5394
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by EATyourGUITAR » Sat Aug 01, 2020 8:03 am

is this an 8bit DAC? that can't be right. it must be 16bit right? maybe this is 10 samples at 16 bit?
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 10:52 am

EATyourGUITAR wrote:
Sat Aug 01, 2020 8:03 am
is this an 8bit DAC? that can't be right. it must be 16bit right? maybe this is 10 samples at 16 bit?
It is 16 bit, these are just the first 8 bits (I only have an 8 channel logic analyzer...)

Also, this is with no additional input signal. There's always data/sound entering/exiting the DAC though, from whatever is going wrong.

User avatar
EATyourGUITAR
has no life
Posts: 5394
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by EATyourGUITAR » Sat Aug 01, 2020 11:49 am

that could be normal for signed numbers depending on the where is LSB. negative counting starts at 11111111 going down with LSB changing to 01111111. positive numbers start at true zero 00000000 then count up 10000000 when LSB changes. this is why overflow of the smallest negative number wraps around and continues counting up through zero. we don't even need to know if it is floating point. we can assume it is signed and we can assume LSB on the left. we can assume that this is the mantissa if it is floating point but it won't matter if you have 16 bits and you only look at the lower 8. there is equivalence there if we only consider these sequences as sequential ordered lists. having everything in floating point changes the encoding and the madnitude but it does not scramble LSB. everything is ordered sequentially either way for the lower 10 bits in 16 bit floating point and also true for all 16bits if it is signed int.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 12:10 pm

EATyourGUITAR wrote:
Sat Aug 01, 2020 11:49 am
that could be normal for signed numbers depending on the where is LSB. negative counting starts at 11111111 going down with LSB changing to 01111111. positive numbers start at true zero 00000000 then count up 10000000 when LSB changes. this is why overflow of the smallest negative number wraps around and continues counting up through zero. we don't even need to know if it is floating point. we can assume it is signed and we can assume LSB on the left. we can assume that this is the mantissa if it is floating point but it won't matter if you have 16 bits and you only look at the lower 8. there is equivalence there if we only consider these sequences as sequential ordered lists. having everything in floating point changes the encoding and the madnitude but it does not scramble LSB. everything is ordered sequentially either way for the lower 10 bits in 16 bit floating point and also true for all 16bits if it is signed int.
Interesting stuff, thanks. The DAC is a PCM54HP, which I'm pretty sure is fixed point. The "overflow" light is always on, which is controlled by the DSP chip, so it would seem like whatever is going wrong, if it's a digital problem, is before the DAC. But I could be totally wrong!

The service manual, observing, and a few anecdotes on the internet seem to narrow it down to the analog switch, the DSP chip, and/or one of the 5 RAM chips. All I've done so far in the RAM world is see that the fast RAM is toggling and that the correct clock signal is going to the logic chips.

The analog switch is finally arriving today so I can eliminate that as a possibility soon.

User avatar
EATyourGUITAR
has no life
Posts: 5394
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by EATyourGUITAR » Sat Aug 01, 2020 12:38 pm

Analog overvoltage indicator light is not the same thing as overflow counting in binary. There is no relation there. You are correct, if the meter says you have a hot signal coming in before that DAC then it is not a problem with the DAC. It could even be square waves if that is all MSB getting flipped. Maybe on that DAC the voltage range only uses 15 bits. It can happen. So either those bits represent a very very small noise with a period 4x your sample rate or it is a massive distorted square wave maxed out with a period about 4x your sample rate. You should probably look before the level indicator. The level indicator would suggest that it is loud distorted square waves going into the DAC.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
Red Electric Rainbow
Super Deluxe Wiggler
Posts: 1736
Joined: Tue Sep 21, 2010 7:48 am
Location: Chicago

Re: Broken Midiverb II: Troubleshooting Tips?

Post by Red Electric Rainbow » Sat Aug 01, 2020 12:44 pm

pitch it and get another one 😂
TOO FAR GONE

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 12:51 pm

Red Electric Rainbow wrote:
Sat Aug 01, 2020 12:44 pm
pitch it and get another one 😂
Haha, that's probably the best advice. If I wasn't trying to fix it as a "learning experience" it would be long gone!

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sat Aug 01, 2020 1:58 pm

remember that every other sample of the DAC will be to match the input signal (ADC mode), and then the other sample is to create the output signal.

if the ram chips are swappable, you can take one out at a time and see what happens. if you could figure out which ones stored the high bits, and which stored the low bits (if its organized that way), you can try to keep working chips in the high bit positions.
openmusiclabs.com

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 2:42 pm

guest wrote:
Sat Aug 01, 2020 1:58 pm
remember that every other sample of the DAC will be to match the input signal (ADC mode), and then the other sample is to create the output signal.

if the ram chips are swappable, you can take one out at a time and see what happens. if you could figure out which ones stored the high bits, and which stored the low bits (if its organized that way), you can try to keep working chips in the high bit positions.
Thanks, I'll have to go back and re-read the DAC datasheet and other stuff to understand this stuff better...

I removed each of the RAM chips from the 1st to the 4th, and after the 4th was removed, the "overflow" light turned off for all the programs and the signal passing through the output was just low level noise.

I put them back in one at a time starting with the 4th position (using the chip previously in the 1st position) and after the last chip was installed it was back to overflow light on, crazy noise.

I should record the sound with no input, it really sounds like something is feeding back, and there shouldn't be anything feeding back, I don't think. Which I guess could only mean the analog switch or the DSP is malfunctioning?

User avatar
gloamtrotter
Common Wiggler
Posts: 134
Joined: Tue Nov 19, 2019 4:33 am
Location: NY

Re: Broken Midiverb II: Troubleshooting Tips?

Post by gloamtrotter » Sat Aug 01, 2020 4:57 pm

Red Electric Rainbow wrote:
Sat Aug 01, 2020 12:44 pm
pitch it and get another one 😂
Please no, I’ve been following this thread since the beginning and am emotionally invested at this point

User avatar
Red Electric Rainbow
Super Deluxe Wiggler
Posts: 1736
Joined: Tue Sep 21, 2010 7:48 am
Location: Chicago

Re: Broken Midiverb II: Troubleshooting Tips?

Post by Red Electric Rainbow » Sat Aug 01, 2020 5:44 pm

hahahahahaha
TOO FAR GONE

User avatar
3hands
Super Deluxe Wiggler
Posts: 1124
Joined: Mon Nov 19, 2018 6:36 pm
Location: Ottawa Ontario

Re: Broken Midiverb II: Troubleshooting Tips?

Post by 3hands » Sat Aug 01, 2020 6:03 pm

gloamtrotter wrote:
Sat Aug 01, 2020 4:57 pm
Red Electric Rainbow wrote:
Sat Aug 01, 2020 12:44 pm
pitch it and get another one 😂
Please no, I’ve been following this thread since the beginning and am emotionally invested at this point

Agreed. I want to see this midiverb back doing amazing things. Don’t bin it!! :)
Gum is fun, but not on a cat.

My minds an art gallery.

User avatar
EATyourGUITAR
has no life
Posts: 5394
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by EATyourGUITAR » Sat Aug 01, 2020 8:30 pm

I want you to ground the input of the DAC and record a big dump like 500 bytes. Even if you only have 8 bytes. Grounding it and looking at the data can tell a bit more. You can also move the probes to get the lower 8 bits in a second test.

EDIT: I think you may want to cut a trace so you don't over current your LM311. pin 49 on the ADC should be grounded during the test. do it clean so you can bridge the cut layer with solder. you can cover it with conformal coating later.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sat Aug 01, 2020 9:38 pm

have you run the "self test" procedure as listed in the service manual?
openmusiclabs.com

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 10:18 pm

3hands wrote:
Sat Aug 01, 2020 6:03 pm
gloamtrotter wrote:
Sat Aug 01, 2020 4:57 pm

Please no, I’ve been following this thread since the beginning and am emotionally invested at this point

Agreed. I want to see this midiverb back doing amazing things. Don’t bin it!! :)
Haha I feel invigorated now! Not giving up yet!
guest wrote:
Sat Aug 01, 2020 9:38 pm
have you run the "self test" procedure as listed in the service manual?
Unfortunately it looks like the 2 doesn't have a self test like the 3. That would make things easier... I tried a bunch of button combinations + power on to see if there was one, but no luck.
EATyourGUITAR wrote:
Sat Aug 01, 2020 8:30 pm
I want you to ground the input of the DAC and record a big dump like 500 bytes. Even if you only have 8 bytes. Grounding it and looking at the data can tell a bit more. You can also move the probes to get the lower 8 bits in a second test.

EDIT: I think you may want to cut a trace so you don't over current your LM311. pin 49 on the ADC should be grounded during the test. do it clean so you can bridge the cut layer with solder. you can cover it with conformal coating later.
This is looking a little scary haha. One thing clear in the service manual is if you make a mistake between the comparator output and pin 49, the DSP dies. Would having an oscilloscope on hand provide some kind of alternative test at a different point in the circuit?

The analog switch arrived today, so I can go forward with replacing that I guess and see what happens.

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sat Aug 01, 2020 10:37 pm

what is the part number on the memory ICs? im trying to figure out if they dividing the memory up by time or by bit depth (ie is each SRAM chip 4b of the total 16b address space, or do they do sequential grabs from each chip to get a total of 16b). the V3 seems to only have one ic for the sram.

also, i forgot that its stereo, so the DAC signals are actually:

L out
R out
MSB SAR
... (14 middle bits)
LSB SAR

so you should see a sequential changing of the subsequent bit while in SAR mode.
openmusiclabs.com

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sat Aug 01, 2020 10:49 pm

typical SAR would look like this:

10000000
11000000
10100000
10010000
10011000
10010100
10010110
10010111
10010110

basically, it increments each lower bit, and if there is a match, it holds it. for no input signal, the values will be near 1000000 or 011111111 as this is the center value in the span of 00000000 -> 111111111.

if youre just looking at the top 8b, then there should be a string of 8 where those dont change, because the bottom 8 are currently being toggled to check for their values. you will need to sample faster, though, as the toggle rate is around 300ns. so sample at least at 4MHz.

[EDIT] looking at the datasheet again, its 666kHz on the DAC clock, so you can probably get away with sampling at that rate for short bursts, but maybe safer to do 2MHZ. L and R samples are held for 2 sample clocks, with a small delay between them.
Last edited by guest on Sat Aug 01, 2020 11:00 pm, edited 1 time in total.
openmusiclabs.com

User avatar
EATyourGUITAR
has no life
Posts: 5394
Joined: Tue Aug 31, 2010 12:24 am
Location: Providence, RI, USA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by EATyourGUITAR » Sat Aug 01, 2020 10:59 pm

if this is showing the exposed logic of a successive approximation register then you will need to look at 16 words to get the one word that is the result of the scanning. the exact list of bits scrolling by depends on the final value that matches all the comparator tests inside the ADC. the scan procedure starts with the general case then bifurcates continuously till it arrives at the final DAC value that matches the voltage at the input of the ADC.
WWW.EATYOURGUITAR.COM <---- MY DIY STUFF

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sat Aug 01, 2020 11:26 pm

guest wrote:
Sat Aug 01, 2020 10:37 pm
what is the part number on the memory ICs? im trying to figure out if they dividing the memory up by time or by bit depth (ie is each SRAM chip 4b of the total 16b address space, or do they do sequential grabs from each chip to get a total of 16b). the V3 seems to only have one ic for the sram.
The SRAM, which I think is the same as the MV3, is CY7C128:
https://www.datasheets360.com/pdf/7434058514295188780

The DRAM is 4 x MB81464-12, which might be different than the MV3:
https://pdf1.alldatasheet.com/datasheet ... 64-12.html

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sun Aug 02, 2020 4:06 pm

I attached two .txt files, one is the DAC data input 1-8, the other is DAC data input 1-3 and 16 (just in case seeing the LSB helps somehow...) Both captured at 4MHz with no additional input signal.

Doesn't seem to look like it should? I guess I'm looking to see if the road to the problem leads to the DSP chip or the RAM, presuming the DAC isn't the problem?

Also, just to confirm, the output of the DAC sounds the same as the output of the LM311 feeding the ADC.
You do not have the required permissions to view the files attached to this post.

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sun Aug 02, 2020 5:24 pm

did swapping the analog switch change anything?

the dac data looks ok, im guessing a DRAM chip.
openmusiclabs.com

User avatar
wadesey
Common Wiggler
Posts: 117
Joined: Mon Oct 10, 2011 3:01 pm
Location: Long Beach, CA

Re: Broken Midiverb II: Troubleshooting Tips?

Post by wadesey » Sun Aug 02, 2020 5:52 pm

Thanks for checking that out, I guess I'll swap out the analog switch now and see what happens.

The MV3 manual has a lot of info about how the SRAM is used, but I don't think anything about the DRAM unfortunately.

User avatar
guest
Super Deluxe Wiggler
Posts: 4326
Joined: Mon Aug 19, 2013 11:49 am

Re: Broken Midiverb II: Troubleshooting Tips?

Post by guest » Sun Aug 02, 2020 6:58 pm

ive had a closer look at the data, and there is something odd about the SAR. im not sure if its some algorithm to pick the most probable starting point, but the 2nd bit in doesnt flip sometimes, it gets skipped in the sequence. that could be a number of things, but i would generally assume its not the DRAM.
openmusiclabs.com

Post Reply

Return to “Music Tech DIY”