USB Logic Analyzers

[TheNewLab]

I'd like to get a logic analyzer for use with some retrocomputing projects that use a bus speed of 4 - 16 Mhz.

Scott

Heh. This is where I'm at as well. Mostly interested in retro-computing projects using 8-bit processors like the Z80, etc.  In the end, I decided to buy a 136 channel HP 1660A logic analyzer.  It's way overkill for what I'm doing, but, well.. TEA syndrome and all that.  I also have a Zeroplus LAP-C 16-channel USB device, and a Sparkfun Bus Pirate, so I figure I'm covered one way or another.

The Zeroplus works with Sigrok, which is the main reason I went with that. (I'm something of an ideologue for F/OSS).

Anyway, I'll just say, be wary of Ebay... TEA is no joke.      I already don't know where the heck I'm going to put this 1660A when it comes in.  I'll probably have to throw my TV out to make room for it...

LOL, aside from consuming space, it gets expensive doesn't it.  LOL

[AngusBeef]

Young guy here just getting into electronics. Got a DSLogic Basic and now I'm looking to see if I can upgrade it.

I noticed that only the SRAM is missing when comparing my chip to the photos of PCBs on Sigrok's website. Then I found this thread - so it seems it's a simple as soldering on the SRAM and renaming the firmware files so it downloads a different firmware into the EEPROM...?

I picked up a MT48LC16M16A2P-6A:G SRAM chip (https://www.mouser.com/ProductDetail/340-128308-TRAY) which is different than the one mentioned in this thread (-GT:R I think), but mirrors the chip from the DSLogic Plus hardware photos on SigRok's site.

Also, one gentleman mentioned removing a SEEP and altering a memory address and I don't quite understand what that means. I assume it's the EEPROM?

[pigrew]

Young guy here just getting into electronics. Got a DSLogic Basic and now I'm looking to see if I can upgrade it.

I noticed that only the SRAM is missing when comparing my chip to the photos of PCBs on Sigrok's website. Then I found this thread - so it seems it's a simple as soldering on the SRAM and renaming the firmware files so it downloads a different firmware into the EEPROM...?

I picked up a MT48LC16M16A2P-6A:G SRAM chip (https://www.mouser.com/ProductDetail/340-128308-TRAY) which is different than the one mentioned in this thread (-GT:R I think), but mirrors the chip from the DSLogic Plus hardware photos on SigRok's site.

Also, one gentleman mentioned removing a SEEP and altering a memory address and I don't quite understand what that means. I assume it's the EEPROM?

Yes, I think that's what it must be. I found https://weekly-geekly.github.io/articles/445024/index.html which has photo instructions.... That post links back to this thread... weird.

I ordered a "plus" about a week ago, which I'm excepting to arrive soon. It's possible that the company will implement serial number checks in future firmware releases or something to disable an "upgraded" unit.... we'll see.

I had a misconception about its adjustable thershold. I figured the plus model would use comparators on its inputs, but it doesn't. It just uses the internal threshold voltage of the FPGA as a threshold reference, probably with a PWM DAC from the FPGA to control it. I'm also curious about the input series resistors on the board, as if they are high enough they will seriously reduce the bandwidth since they don't have parallel capacitors. I wonder if its response could be improved by putting some ~47 pF (wild guess) caps on top of the resistors on each input channel (the column of resistors right at the FPGA). Also, is the FPGA using 3.3V or 1.8 V logic levels on the input? Probably the 1.8 would give a more reproducible threshold.

So, I've never used the DSLogic before, but it does seem like it's a not too difficult thing to do.

Another complaint is that the probe leads are too long! They should be shorter to minimize the distance of their compensation network to the probe. I'd be curious about measuring their input impedance. However, adding an additional board in the middle of the probe would add cost, and maybe not increase the performance by much.

[AngusBeef]

Yes, I think that's what it must be. I found https://weekly-geekly.github.io/articles/445024/index.html which has photo instructions.... That post links back to this thread... weird.

I ordered a "plus" about a week ago, which I'm excepting to arrive soon. It's possible that the company will implement serial number checks in future firmware releases or something to disable an "upgraded" unit.... we'll see.

I had a misconception about its adjustable thershold. I figured the plus model would use comparators on its inputs, but it doesn't. It just uses the internal threshold voltage of the FPGA as a threshold reference, probably with a PWM DAC from the FPGA to control it. I'm also curious about the input series resistors on the board, as if they are high enough they will seriously reduce the bandwidth since they don't have parallel capacitors. I wonder if its response could be improved by putting some ~47 pF (wild guess) caps on top of the resistors on each input channel (the column of resistors right at the FPGA). Also, is the FPGA using 3.3V or 1.8 V logic levels on the input? Probably the 1.8 would give a more reproducible threshold.

So, I've never used the DSLogic before, but it does seem like it's a not too difficult thing to do.

Another complaint is that the probe leads are too long! They should be shorter to minimize the distance of their compensation network to the probe. I'd be curious about measuring their input impedance. However, adding an additional board in the middle of the probe would add cost, and maybe not increase the performance by much.

I've watched one of the videos above and it appears that the way they designed the input leads and the input series resistors leads to ringing in circuits and places a significant load on the DUT. But they recommend it anyway so it may not be enough to trigger the threshold values?

Also, I think serial number checks are unlikely - DSLogic is mostly defunct now from everything I've heard. Even their website still says 2018 at the bottom. And there are so many clones.

DSLogicBasic-

Soldered on a MT48LC16M16A2TG-6A:GTR ($2.33)

Upgrade DslogicBasic to DslogicPlus !!!!!

I did the above  thing, plus reprogramming the SEEP 24LC128 .
Take out the SEEP , Read and its data to file. edit this files , search for exists of data "0E2A2100".
Modify this data to "0E2A2000". and save file.   Write this data file to SEEP.
It works................Most importantly, no need to fool around with the firmware files..

I would guess this gentleman is the one who wrote the article you linked. I have no idea how to program EEPROM since I've really only been learning electronics for a month so it'll be an adventure. I would think it's possible to flash it through the USB interface, but I have no idea how to do that either. Yet!

[pigrew]

I've watched one of the videos above and it appears that the way they designed the input leads and the input series resistors leads to ringing in circuits and places a significant load on the DUT. But they recommend it anyway so it may not be enough to trigger the threshold values?

Also, I think serial number checks are unlikely - DSLogic is mostly defunct now from everything I've heard. Even their website still says 2018 at the bottom. And there are so many clones.

I would guess this gentleman is the one who wrote the article you linked. I have no idea how to program EEPROM since I've really only been learning electronics for a month so it'll be an adventure. I would think it's possible to flash it through the USB interface, but I have no idea how to do that either. Yet!

I think that the bandwidth is probably good enough, up to a point.... My guess is 50 or 100 MHz bandwidth, not sure. I don't even know how I would define the bandwidth of the LA.... I hope the bandwidth is at least close to the sample rate, but I doubt that it is.

It looks like just swapping the firmware filenames or editing the profile list would be the easiest option. I think that it's serial flash, so you could write to it from a mcu, or buy a special programmer to write to it. But, it's easiest to just try swapping the files on the computer.

I don't think that DSLogic is quite dead. They just released a new version of their DSView software. looking at it, it has a lot of room for improvement, but it's usable. There are a huge number of branches on GitHub, but few if any ever get merged back into the trunk. I don't think that there are any clones, I think that they're just selling through the "normal" online Chinese distribution networks where everyone sells it on TaoBao without any sellers knowing any details about it......

The length matching of the tracks on the PCB is beautiful. I'm curious about how well it works with its "use external clock" mode. Do they trigger the IO cell registers on the external clock, or do they fake it by just analyzing the data they sampled at 400 MHz? How much skew is there?

[pgo]

You should look at this thread for how to program over USB without requiring a programmer.
Requires soldering a link however.

https://www.eevblog.com/forum/testgear/upgrading-dslogic-basic-to-plus-without-eeprom-modification/msg2594313/#msg2594313

[AngusBeef]

You should look at this thread for how to program over USB without requiring a programmer.
Requires soldering a link however.

https://www.eevblog.com/forum/testgear/upgrading-dslogic-basic-to-plus-without-eeprom-modification/msg2594313/#msg2594313

You don't have to solder the link, just connect pins 4 and 7 with two EZ-Hook style probes.

You do have to solder the RAM chip, but it's simple drag soldering.

[Crazy_Fox]

The length matching of the tracks on the PCB is beautiful.

Could you explain, please, how do you measured the length of tracks, or just say what the length is ?
I got U2Basic version, and on the PCB they removed track from FPGA to A12 pin of DRAM, so seems I need to add the wire link with the same length as tracks to be able to use 256Mb.

[pgo]

Hi Crazy_Fox,

I'm surprised they have done this.  I checked on the one I have and it has a connection to A12.  I suggest you check this yourself if you have not done so.

You can try measuring the traces on a photograph and a high zoom and then scale appropriately.
Suitable photos are available on this thread: https://www.eevblog.com/forum/testgear/usb-logic-analyzers/msg1333915/#msg1333915

Note that the track is below the IC and then on the rear of the board with two vias.

bye

 

[liyang]

I got U2Basic version, and on the PCB they removed track from FPGA to A12 pin of DRAM, so seems I need to add the wire link with the same length as tracks to be able to use 256Mb.

I'm surprised they have done this.  I checked on the one I have and it has a connection to A12.  I suggest you check this yourself if you have not done so.
…
Note that the track is below the IC and then on the rear of the board with two vias.

Hi Crazy_Fox & pgo,

I just got my U2Basic (unfortunately) from Banggood, and it looks like I don't have A12 connected either. (There's no via on mine compared to the photo pgo mentioned.)

@Crazy_Fox: my board is labelled DSLogic-U2B-V111-P0808; what's yours?

[jake111]

I'd like to get a logic analyzer for use with some retrocomputing projects that use a bus speed of 4 - 16 Mhz. Checking ebay, I can find the following:

* Salae 8-channel 24 Mhz for around $6
* Salae 16-channel 100 MHz for around $43
* SeeedStudio 16-channel 100 MHz / 4-channel 400 Mhz for around $75

I'm thinking the 16-channel 100 Mhz products fit my use case with room to expand. The 8-channel 24 Mhz is appealing solely because at $6 if I do something stupid and ruin it, I can throw it away and not feel bad.

Any experiences with these products? is the software decent?

Scott

I have used a great many logic analyzers over the years in troubleshooting digital bus issues the origin of which were in the analog domain.  I absolutely love the Saleae but I have the pro series that lets you stream through USB 3.0 and this is a huge benefit.  Saleae's software is really great.  If you are always able to set up creative triggers for what you are working on, then you can make the cheaper ones useful, but I have used quite a few of these chinese logic analyzers that are really just copies of copies of copies and as you would expect, they are all kind of junk in one way or another.  I like products that are designed to function well for their purpose, not products which are just copies of other designs optimized to be as cheap as possible and assembled or packaged with child labor.