Author Topic: Seeing ramps in a digital signal  (Read 1288 times)

0 Members and 1 Guest are viewing this topic.

Offline vampi-the-frog

  • Regular Contributor
  • *
  • Posts: 50
  • Country: ro
Seeing ramps in a digital signal
« on: March 30, 2017, 10:18:36 am »
Hello friends. I am trying to fix a faulty Korg M1. I've taken out the main board, and poked around with a scope. I am seeing signals like the ones in the attached scope screenshot. The signal is measured on the 74LS245. Top cyan signal is OE# (output enable, negated). Second, yellow signal is DIR (direction), third, purple signal is B7 and the last, blue signal is A7. It could be A6 and B6. It doesn't matter, on some pin pairs this happens, on some it doesn't. As you can see from the attached schematic snippet, it is an input/output buffer on the CPU data/address bus. I have literally desoldered all the ICs from the board that would use the data bus, and now it's just the CPU, the 74LS245, and some misc chips that should not interfere. I am seeing these ramps, sort of like a capacitor charging, perhaps, but I am stumped as to where they come from. I've replaced both the 74LS245 with new ones. There is a resistor array also pulling up the signal aparently, but I've tested those and they're all 10k. It's just the CPU and the buffer now. It might be the CPU, I am not sure though. This board is a 5V board that possibly got fried due to the power supply going up to 7V. In practice, it works intermittently. The LCD displays text, sometimes with garbage characters, but the keyboard itself makes no sound. I did get it to make sound at one point, and at another point it was sending MIDI data (the MIDI UART is inside the CPU, connected pretty much directly to it). I do not understand where the ramps come from, I am pretty sure it's supposed to be a clean square signal, not ramping. Do you guys have any suggestions please? Thanks.


« Last Edit: March 30, 2017, 10:24:09 am by vampi-the-frog »
 

Offline vampi-the-frog

  • Regular Contributor
  • *
  • Posts: 50
  • Country: ro
Re: Seeing ramps in a digital signal
« Reply #1 on: March 30, 2017, 10:30:21 am »
More screen captures. Sometimes it looks like shark fins.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1470
  • Country: us
Re: Seeing ramps in a digital signal
« Reply #2 on: March 30, 2017, 11:09:32 am »
Sure looks like capacitance to me too.

What probes are you using?  If they're x1/x10 switchable, are you using your probes on x1?  If so, use x10.  You're adding a lot of extra capacitance when you use the x1 setting.  You're going to see some ramp even on x10 with a 10k pullup, but not quite so long.

It's hard to tell, but it looks like you have something funny going on in your first screen capture on the Ax line.  The signal looks like it's getting above +5V, but it could be an accuracy issue at that resolution.
 

Offline Alex Eisenhut

  • Super Contributor
  • ***
  • Posts: 2034
  • Country: ca
  • If you can buy it for 4$ on eBay, why design it?
Re: Seeing ramps in a digital signal
« Reply #3 on: March 30, 2017, 11:16:29 am »
Looks like a perfectly normal three-state bus to me.  :-//

As long as Vil and Vih are respected, who cares? When the bus floats, pull ups will exhibit a perfectly normal RC curve.

You'll benefit greatly from turning off the scope, especially the soldering iron, and read up on three-state logic, and hopefully a service manual or data sheet...
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1470
  • Country: us
Re: Seeing ramps in a digital signal
« Reply #4 on: March 30, 2017, 11:31:37 am »
Looks like a perfectly normal three-state bus to me.  :-//

As long as Vil and Vih are respected, who cares? When the bus floats, pull ups will exhibit a perfectly normal RC curve.
Yes, but look at the time scale.

500ns to get to about 60% of 5V with a 10k pullup comes out to be about 100pF.  That's too much.
 

Offline Alex Eisenhut

  • Super Contributor
  • ***
  • Posts: 2034
  • Country: ca
  • If you can buy it for 4$ on eBay, why design it?
Re: Seeing ramps in a digital signal
« Reply #5 on: March 30, 2017, 11:52:47 am »
It's broken. He desoldered all the chips so these waveforms are useless garbage now. And said it was powered with 7V. Who knows how much more damage was done by the desoldering?

Without a datasheet of the processor, who really knows what timescale is good or not?

Put a scope on a random bus signal on contemporary 1980s digital electronics like an Amiga and you'll see plenty of nonsense like this.

It means nothing without some context, I'll stick to my advice. Nowhere did OP mention "tri-state", I'll assume he doesn't know what that is.

My usual approach for troubleshooting complex digital monsters is always the same.
1) Power supply. Test it with a dummy load. Use a DMM. Read the manual/datasheet to get an idea of the current it's supposed to be able to deliver.
2) Reset line. It's surprising but without a clean, proper reset signal that correctly covers the power supply ramp up and crystal stabilization, bad things can happen. Sometimes the reset comes from a 555-type circuit that uses a small electrolytic cap in the RC circuit. It goes bad, the reset goes bad.
3) Clock. There needs to be a correct, stable clock somewhere. This gets tricky when it's built-in to the chip, as probing a high-impedance crystal oscillator usually just kills the oscillation. Find a convenient source of buffered signal and check that.
4) Get as much information as possible from other people! It's a fairly popular instrument, there must be tons of troubleshooting info out there.

The last thing you do is probe one signal, then proceed to stripping the thing down to the fiberglass!

100pF? Like a bus loaded with 8 or 16 DRAM chips and dozens of jellybeans? It's not out of the question. 100uF would be bad, yes.

OP: What do you think OE does?
 

Offline pieman103021

  • Contributor
  • Posts: 38
  • Country: us
Re: Seeing ramps in a digital signal
« Reply #6 on: March 30, 2017, 12:38:21 pm »

The last thing you do is probe one signal, then proceed to stripping the thing down to the fiberglass!


I'm still trying to understand how OP came to this as a solution. There are maker/hacker communities devoted to korg products and even more information related to their repair, there is no reason to do anything so severe without consulting them first.
Sorry, don't mean to be roasting you OP :-\. Hoping you are able to get this fixed.
Exclamation Point!
 

Offline Armadillo

  • Super Contributor
  • ***
  • Posts: 1725
  • Country: 00
Re: Seeing ramps in a digital signal
« Reply #7 on: March 30, 2017, 01:54:55 pm »
Use a cheaper scope or limit the frequency of the scope or use average. Read the oscilloscope manual.
 

Offline vampi-the-frog

  • Regular Contributor
  • *
  • Posts: 50
  • Country: ro
Re: Seeing ramps in a digital signal
« Reply #8 on: March 30, 2017, 08:52:41 pm »
You'll benefit greatly from turning off the scope, especially the soldering iron

It's broken. He desoldered all the chips so these waveforms are useless garbage now. And said it was powered with 7V. Who knows how much more damage was done by the desoldering?

.
.
.

OP: What do you think OE does?

Electronics seems to be an alchemical mix of technology and emotions. People like to take their hidden anger out on beginners for some reason, and here is my answer to your anger: nobody's forcing you to reply, and it's not my fault that you're pissed. I'm posting in this international forum for the exact reason that on my local electronics forum everyone's quite salty.

Anyway, moving on.

Sure looks like capacitance to me too.

What probes are you using?  If they're x1/x10 switchable, are you using your probes on x1?  If so, use x10.  You're adding a lot of extra capacitance when you use the x1 setting.  You're going to see some ramp even on x10 with a 10k pullup, but not quite so long.

It's hard to tell, but it looks like you have something funny going on in your first screen capture on the Ax line.  The signal looks like it's getting above +5V, but it could be an accuracy issue at that resolution.

I was using them on 10x, all 4 of them, I checked. There is extra ringing due to the fact that I hadn't grounded two of the probes, it went away after grounding, however, the problem you mentioned I've observed previously: The Ax line sometimes goes to 4V and sometimes to 5V.

Someone elsewhere has suggested that there is some capacitance in the FETs in the ICs.

Also, if it is the probes, then switching to x1 should produce a different result, correct? I'll try that.

I think indeed that I need to figure out how this 3-state business works. I'll have a better look.

Without a datasheet of the processor, who really knows what timescale is good or not?
It's a NEC  ?PD70216GF-8, part of the NEC V50 cpu family.

It means nothing without some context
Here is a scan of the service manual. I have desoldered all the chips on the data bus that comes out of the buffers.

My usual approach for troubleshooting complex digital monsters is always the same.
1) Power supply. Test it with a dummy load. Use a DMM. Read the manual/datasheet to get an idea of the current it's supposed to be able to deliver.
2) Reset line. It's surprising but without a clean, proper reset signal that correctly covers the power supply ramp up and crystal stabilization, bad things can happen. Sometimes the reset comes from a 555-type circuit that uses a small electrolytic cap in the RC circuit. It goes bad, the reset goes bad.
3) Clock. There needs to be a correct, stable clock somewhere. This gets tricky when it's built-in to the chip, as probing a high-impedance crystal oscillator usually just kills the oscillation. Find a convenient source of buffered signal and check that.
Good advice. I had tested the power supply without a load. I'll try to test it under load of some sort. It seems to regulate the plus and minus 12 volt rail relative to the 5 volt rail, so I guess only the 5 volt rail needs to be loaded for testing.

Also, I hadn't thought of testing the reset circuit. The clock seems fine, 32MHz.

4) Get as much information as possible from other people! It's a fairly popular instrument, there must be tons of troubleshooting info out there.
I've consulted the service manual. I am not sure how common the problem is.

There are maker/hacker communities devoted to korg products and even more information related to their repair, there is no reason to do anything so severe without consulting them first.
Could you link me to a couple of them please? I haven't been able to find them.

Thanks for all your replies guys!
 

Offline pieman103021

  • Contributor
  • Posts: 38
  • Country: us
Re: Seeing ramps in a digital signal
« Reply #9 on: March 31, 2017, 12:33:47 am »
You can find information on their repair by looking up information on vintage synthesizers or Korg repair. You will also find forums where you can ask some questions.
I don't really know anything about music stuff, and I don't want to make a recommendation based on my lack of knowledge.
Exclamation Point!
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1470
  • Country: us
Re: Seeing ramps in a digital signal
« Reply #10 on: March 31, 2017, 01:11:09 am »
...
I was using them on 10x, all 4 of them, I checked. There is extra ringing due to the fact that I hadn't grounded two of the probes, it went away after grounding, however, the problem you mentioned I've observed previously: The Ax line sometimes goes to 4V and sometimes to 5V.

Someone elsewhere has suggested that there is some capacitance in the FETs in the ICs.

Also, if it is the probes, then switching to x1 should produce a different result, correct? I'll try that.
...
No, don't use x1 on high speed circuits.  x10 is correct.

I thought you might have been using x1 which would have added extra capacitance to the circuit and contributed to the ramp.  That was the main point of my post.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf