Nord Lead 3 problem, does not boot [SOLVED]

[23skidoo]

I would absolutely send it to support, let's see what they say. If they can't fix it and Sweden for whatever reason won't or can't, then we'll look at more invasive actions. But it may be as simple as "whoops that chip somehow got erased during a failed upgrade" or whatever, and they can restore it using a simple re-program. I can't find a supplier for the SST27 series anymore, and you can't drop an SST39 in place (well you theoretically maybe could if you pre-programmed it and then cut the pin 1 connection, basically making it a ROM, but that's just theory from a quick skim and I'm not certain that in practice it would work).

[hozone]

I would absolutely send it to support, let's see what they say. If they can't fix it and Sweden for whatever reason won't or can't, then we'll look at more invasive actions. But it may be as simple as "whoops that chip somehow got erased during a failed upgrade" or whatever, and they can restore it using a simple re-program. I can't find a supplier for the SST27 series anymore, and you can't drop an SST39 in place (well you theoretically maybe could if you pre-programmed it and then cut the pin 1 connection, basically making it a ROM, but that's just theory from a quick skim and I'm not certain that in practice it would work).

The blank part is availablehttps://www.findchips.com/search/sst27sf010. A blank part is not helpful though.
hozone's measurement shows the Vpp pin connects to 5V, so their NL3 lacks the capability to update it.

@hozone Thanks for the photo but I do not see anything obvious to try. I agree with 23skidoo - it's time to send it to Support.

[23skidoo]

Oh nice, thanks for the tip on that part. I might get a few and see if I can't whip an Arduino into programming them. Just in case, you know.

[hozone]

The repair will not be cheap, I'll try myself. I'm not so sure it's the ROM, It may be the SRAM chip, that are cheap and relatively easy to change.

First of all, I've to understaind the behaviour of my RIGOL (see my post here: https://www.eevblog.com/forum/testgear/ ... -question/), it makes me not aquire a good input. It's "strange" to me, of course I'm not a pro
I've acquired images also with the logic analyzer, see attached.
Using the analyser it seems CLKOUT goes from 8Mhz to 5Mhz, but I'm again not sure about this instrument at that speed, I've never use at this frequency.
One again thank you.

[23skidoo]

Well the skips on your LA show me that it's not reading the signal right, if that's a clock signal it should be perfectly regular. When you're dealing with signals in excess of a few MHz you need short traces and good grounding to get a clean signal. You also need to set the LA threshold correctly for the signal being analyzed and the only way to do that is to have worked with it on known signals ahead of time. Troubleshooting a piece of gear without knowing your LA and how it behaves on similar signals is a recipe for confusion.

I assume you're familiar with the Nyquist theorem relating sample rate to bandwidth, right? Bandwidth on an oscilloscope is the -3dB down point from an analogue perspective, so your 50MHz scope is 3dB down at 50Mhz. It has nothing to do with sample RATE which is how fast new samples are acquired, which as pterm mentioned must be more than twice the frequency you wish to measure or you will get aliasing. The same principle applies with your LA - you need to check the sample rate - and in the window of your screenshots here it looks like it is 16MHz which would give you a maximum bandwidth of 8.0MHz. Anything over that will alias. Note that digital pulses (square waves) typically need about 3-10x the bandwidth depending on how precisely you need to see the edges, so if you want to properly see a pulse on a logic analyzer you need AT LEAST 3x the bandwidth of the signal you want to measure and that's cutting it very close by most lab definitions. I prefer to have closer to 5x as a low threshold.

From the 3x perspective you'd need >50MHz sample rate to observe the 8.389MHz clock correctly, and a 150MHz sample rate to observe a 25MHz clock or digital datastream correctly. Anything lower and you won't be seeing the true signal.

For instance, when dealing with 50MHz SPI signals, I use a Rigol MSO5104 which has up to an 8Gsa/sec sample rate (spread across all the active channels; with 16 active channels that still leaves me 500MHz of sample rate and 250MHz of bandwidth, 5x the 50MHz signal I want to observe). That's about the minimum speed I'd use for signals that high. Remember for scopes and logic analyzers that (in all equipment I'm aware of and have used, at least) you must divide the current "scope sample rate" by the number of active channels to get the channel sample rate. Thinking about it, you had 2 channels active on your LA, so perhaps you only had a true per-channel BW of 4MHz... hmmmm. Again, the only way to know for sure is to use it on real, known, signals first and see how it behaves.

[hozone]

Thank you @23skidoo,
I've miss your reply, sorry about that.
I'm a IT developer, so I'm not a pro in electronic, that's just hobby for me. Thios is the reason of my mistakes. Anyway I really approciate your help. I know the Nyquist, but forgot about that Also, the 3x rule o thumb will help me.
I'll keep you updated as soon as I get time to work on this synth again. Unluckily my job takes me most of my time those period.
Again, thanks!

[hozone]

It's alive! THANK YOU ALL!
I finally had the time to replace the components I've buy, and the broken thing was...
...
suspense...
...
one of the SRAM CY7C109B IC, It was not the first one I've replaced, It was the second one, but I was lucky!
It costs me a few hours and couple of Euros.
The most important thing is that I've learn something thanks to you, and also I've salvaged this keyboard from his end.

[hozone]

Excellent!

[hozone]

Good job!

[Hlaalu]

It's good to know there's such a wonderful community willing to help when needed.