Help with Linn LM-1 Drum Computer.

[FrankieMusic]

Hi guys. I have a problem with my Linn LM-1 Drum Computer. The conga drums have a high-frequency distortion in them that makes them sound brittle, almost metallic, almost like cowbells. I sent it to Bruce Forat a few years ago, but he was unable to hear what I was talking about.

I've taken some dry samples (from the stereo output and the individual out for the low conga) for you to listen to and hopefully identify what could be wrong. You can find them here:

http://www.fileden.com/files/2009/4/7/2 ... Congas.wav

http://www.fileden.com/files/2009/4/7/2 ... vidual.wav

Please listen to them and let me know if you have an idea of what could be wrong. Thanks!

[tim gueguen]

I think you're just hearing the limits of the sampling technology available at the time. They're only 8 bit and 27 khz.

[FrankieMusic]

I think you're just hearing the limits of the sampling technology available at the time. They're only 8 bit and 27 khz.

I know what the sampling rates are for the sounds, and I strongly doubt that I am hearing the limitations of the sampling technology.

I've compared my samples with a sample of an unfiltered conga from the LM-1 that can be heard at this page:

http://electrongate.com/dmxfiles/linn/index.html

The sample there sounds clear and bright. Mine have a metallic quality to them. Listen to the sample from electrongate and compare it to my samples and then tell me that I am "hearing the limits" of 8-bit sampling technology.

[Overgear]

I hear it and I see it. There is a DC offset in the VCA env( (I think it's a VCA?). Therefore, it clicks at the start and end of each sound and sounds overdriven...because it is. I'm sure you know what DC offset is and, of course, this is internal not a front volume slider. There's gotta be a little pot inside for VCA level. I tried to look at the schems I have but I didn't see with a cursory look.

I confirmed this by recording the outputs of mine. I could see(in a Spectral Analyzer) a big spike around 20hz every time there was a hit in your first audio file. Looking at the wavedata, you can see big jumps at the beginning and end of each hit. Another definitive sign of an VCA DC offset to me. Mine has no sign of this. Nothing at 20Hz.

One of my SEMs has a slight offset right now. Not this bad, though. Still, I should go tweak that pot...

Are the toms the same way? They're on the same drum-board is why I ask.

Hope this helps.

[FrankieMusic]

Overgear, you are a lifesaver!

I haven't the faintest idea of what DC offset is, but your description sounds totally accurate. The congas do sound overdriven; the metallic harshness I've described sounds like soft-clipping.

I don't want to start fooling around with the insides of the machine; I will provide Bruce Forat with this information the next time I send the machine for service.

As far as the toms go, I can hear very subtle clicks at the start of the samples, but they don't sound overdriven like the congas. Still, I will tell Bruce to check the DC offset in the VCA envelope on both the toms and congas.

Is it possible for you to streamline the description of the problem for me so I can describe it accurately to Bruce? I would greatly appreciate it.

Thank you so very much for identifying the problem!

[Overgear]

Here's a good explanation of of DC Offset. With pictures :D . I like pictures. Pictures help with this.

After further review there may be more going on. Looking at your hits and comparing them to mine, I see the offset but it also looks to be cutting the samples off a little early on yours. I have pictures :D . Be sure to copy them if you need to.

Ok, the top diplay is yours and the bottom is mine. I tried to match pitch and level to your first audio file when I recorded mine. Yours is less than a db louder than mine, peak level not RMS. Zoom levels are the same and I lined up a low and hi conga hit. Look at the tails or decays. Yours look chopped yet you still see that really low frequency signal trailing off in its place.

Same files here zoomed in on the hi conga hit at the point where the sample itself stops or gates. Horizontal zoom(time) is equal but vertical zoom(volume) is not. Look at the numbers on the left and you can see yours cuts off almost 20db higher(or louder) and when it does cut off it doesn't go towards the center line(which it should - as mine does). It actually moves away from the center line.

This is a shot of one of the hi conga hits in the 30 band Spectral Analyzer that comes with most RME audio interfaces. This is from your first audio file. Notice the strangely lonesome peak all the way left at 25hz.

This is one from my recording. While there may be overall differences of maybe 5db between yours and mine(say, at 400hz) I don't think that will explain the difference at 25hz. That peak is the sign of a DC offset.

But... looking at the schems and actually opening mine up all I found was a VCF trimmer( I was calling them pots, oops) for the Toms/Congas. I don't know if just the filter alone could cause this problem. The Amplifier(s?) for the Toms/Congas may be discrete or just not tweakable with a trimmer. It may be component failure or the beginnings. Like an op-amp or something.

I don't know if this is streamlined but it may help show there is a problem. Maybe someone else can tell you more from the pics. Hope it helps.

[FrankieMusic]

Thanks, Overgear. I'm sure I can explain the information to Bruce as best I can. Thank you for helping me finally diagnose the problem.

[Overgear]

No prob. Glad I could help.

I have to admit, I did examine this a little more. I think this might be an easier way to describe it.

The DC offset is a side-effect of the problem. The sample being overdriven seems to be the issue. It looks and sounds like it is happening before it gets to the VCF. That is just a guess and an opinion, though.

If I applied a HPF at 20hz to your audio file I could reduce the clicking but it still leaves you with a distorted sound. Not much you can do about that.

Good luck.

[madtheory]

Thanks, Overgear. I'm sure I can explain the information to Bruce as best I can. Thank you for helping me finally diagnose the problem.

Interesting thread. I'm sure you could email it to Bruce...

[FrankieMusic]

Interesting thread. I'm sure you could email it to Bruce...

I will do just that.