Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 19, 2024, 12:59:22 AM

Login with username, password and session length

Recent Posts

 

Heatsink questions on Crate

Started by txflood, February 12, 2013, 12:33:41 PM

Previous topic - Next topic

domohawk

Thank you for the quick reply! Exciting to think some headway may be made on this. I am new to this forum and indeed just getting into electronics hacking! :)

Please bear with my noobness if I am overlooking where the pdf is.. Could you please refer to the pdf with the 47ohm R and bypass cap? They do not appear in this thread, and the closest I can find elsewhere is http://elektrotanya.com/crate_gx_120,_gx_212_sch.pdf/download.html but that does not appear to have j10 dsp connection.

I assume you are referring to R107 being increased 10 > 47 ohms? Is the bypass cap an additional 0.1uf cap put in parallel to R107? (I assume you are not referring to C63 as this is in my schematic/amp as well).

smackoj

#31
I am the Proud owner (what?) of a Crate DT50C that I bought mostly because it came with 2 nice MIE Celestion 10s. It hummed badly and one of the two inputs was inoperable. I replaced the power filter caps and it had little to no effect on hum. Then I did as others have suggested, thorough inspection and re-flowed the solder joints on everything I could get my solder tip close enough to especially around the input jacks, potentiometers and power supply. I also had a Native American medicine man perform an exorcism on the demon spirits that typically haunt Crate equip. Wala, it only hums 1/3 as much and I use it as a backup at church. One great thing about this amp is that I don't worry that someone will steal it and I can torture it with any combination of pedals and gain settings without worrying if it blows up.    :loco

J M Fahey

Well, you did the right thing.
Strong hum and unoperable input jacks point straight at cracking in that area.

Which is very common, it´s very easy to yank a guitar cable in a bad way, or have the amp fall on its front breaking the plug or simply stepping on the cable.

Any/all of these will either break the jack itself or make it twist strongly and crack/break whatever they are soldered to, in this case the PCB.

Roly

Quote from: domohawk on May 20, 2014, 07:14:42 AM
Please bear with my noobness if I am overlooking where the pdf is.. Could you please refer to the pdf with the 47ohm R and bypass cap? They do not appear in this thread, and the closest I can find elsewhere is http://elektrotanya.com/crate_gx_120,_gx_212_sch.pdf/download.html but that does not appear to have j10 dsp connection.

I assume you are referring to R107 being increased 10 > 47 ohms? Is the bypass cap an additional 0.1uf cap put in parallel to R107? (I assume you are not referring to C63 as this is in my schematic/amp as well).

The 47 ohm is the new value for the 10 ohm resistor in series with the supply to the Fx DSP.  The new 0.1uF cap goes between the DSP supply pin and ground at, or near as possible to, the DSP connector.  This should reduce the sing-song digital crud getting out.

Earlier posts suggest that beefing up the earlier electrolytic filter cap for this supply may help the hum problem.

I've had a good search for the previous post of this circuit so I could just link to it, but I can't find it; so...
If you say theory and practice don't agree you haven't applied enough theory.

g1

#34
 Roly, the schem. with the 10 ohm has its own supply, not running off the 40V line.
The gfx212 schem. shows a different supply and 470, not 47 ohm.  Are we looking at the same schematics?

Roly

Quite likely not;

Title: GFX120a
No: 07S252-XX
Sheet: 1 of 2
Date: 6/19/98

The basic point is that one has extra bypassing the other doesn't, and this change in DSP supply arrangement may be a clue they were having some problems in this area.
If you say theory and practice don't agree you haven't applied enough theory.

domohawk

Roly, I am also seeing the 470 ohm.

How exactly would this 'bypass filter' work? Wouldn't the 0.1uF cap just roll off the very high frequencies?  The hum is pretty low, I'd guess around 240hz.

Roly

Let me be the first to say that this is a bit of a shot in the dark.

Quote from: domohawkHow exactly would this 'bypass filter' work?

Bypass filtering of supply lines normally consists of a resistor in series, then a cap to ground.  This forms a First Order low pass filter which has a cutoff frequency where the reactance, Xc = 1/2 Pi f C, of the cap C is equal to the series resistance, Xc = R.  The bigger either C or R are, the lower the frequency.

For 470 ohm and 0.1uF this is;

Xc = 1/ 2 Pi f C

f = 1/ 2 Pi Xc C

1/(2*Pi*470*(0.1*10^-6)) = 3,386.3Hz or 3.4kHz.

This would help prevent high frequency noise/signals getting in to the DSP, and would help to reduce DSP noise getting out, but have no effect on hum on the supply lines.

If we were to place, say, a 100uF of suitable voltage rating across the 0.1uF we would get;

1/(2*Pi*470*(100*10^-6)) = 3.4Hz, well below any mains hum.


"Hum" in amplifiers is by convention power mains related, either 50/60Hz sinewave by direct pickup, what you get when you touch the tip of your guitar lead, or a sawtooth at double that frequency as the result of full wave rectification in the power supply getting into the audio path somehow, generally through insufficient filtering, either due to design error or more commonly due to aging/failing filter caps.  4th harmonic 200/240Hz is pretty unlikely but it is easy to be fooled by 2nd harmonic sawtooth because the waveshape itself has a strong 2nd harmonic 200/240Hz content 'tho its fundamental is actually 100/120Hz.

Yes, 470 ohms and 0.1uf has a high cutoff and is intended to keep the digital noise in the DSP board, but if the DSP board is suffering from hum related to its supply then this suggests that this supply needs close examination.  Assuming we can discount a faulty part* and an earth loop, does this supply need extra filtering, or even regulation in the form of a zener or three-pin regulator?

(* have a very close look at the DSP module with a bright light and lens for any evidence of on-board electros leaking or swelling)

The amp either didn't come out of the factory with a hum-on-DSP problem, in which case a component has failed (e.g. filter electro gone low-C); or it did, in which case we have a basic design problem that can only be fixed by doing the design work that was skipped over, and modifying the amp accordingly.

The changes in that DSP supply area suggest to me that they found a problem after they started their initial production run and modified it on later runs.  While this mod looks like it was to address higher frequencies getting into/out of the DSP it is a bit of a smoking gun now that these amps seem to also have a problem with hum going in via the same route.


So I'd try adding a suitable electro cap across the supply at the DSP board socket.

This will have one of three outcomes;

- Hum eliminated - we are right on the right track.

- Hum reduced - we may be on the right track (try a bigger cap).

- Hum the same - it's something else (it's a red herring).
If you say theory and practice don't agree you haven't applied enough theory.

domohawk

#38
Roly, thank you for your in depth reply! It helped me learn a bit more about filter circuits.

I suspect adding a higher F cap will provide much better results in getting the 120hz out, but I will also give the higher R and 0.1uF a go. Just waiting on digi-key now.. I even picked up an 8200uF cap in addition to the 100uF for fun, since it was less than 2 bucks :)

For the interested, a bit of backstory I think leads to an explanation at why I think this "crate dsp hum" problem is so prevalent on the internet:

In my case I did not notice hum when I bought the amp many years ago. Then one day it got a REALLY LOUD hum, out of the blue. After many hours googling I found a solution was to disconnect the dsp and it was dead quiet.

Fast forward a few years to last week when I decide to take a look at it again and see if I can salvage the dsp use. More relentless googling leads me to the answer, in a 2011 forum post (last: http://music-electronics-forum.com/t13806/). The problem is the crappy pc power connector (J10). The J10 on the dsp is mounted horizontal, so after hours of play the connector rattles loose and causes a huge hum. Crimping the connector or better yet soldering it removed the LOUD hum.

So if the LOUD hum is fixed why am I posting this? Well because the dsp still adds hum (I'd say 20% of the LOUD hum) over no dsp, but since I never knew the crate could be ninja silent with the dsp disconnected, I didn't notice the problem when I first used it. My ears have probably gotten more picky as well.

So in final, if txfloods 'over filter' cap does indeed decrease this 20% hum by 2/3 I will indeed be a happy hacker! Of course I'm still adding the true bypass switch because I'm persnickety :)

g1

Quote from: Roly on May 20, 2014, 02:30:39 PM
Quite likely not;

Title: GFX120a
No: 07S252-XX
Sheet: 1 of 2
Date: 6/19/98

The basic point is that one has extra bypassing the other doesn't, and this change in DSP supply arrangement may be a clue they were having some problems in this area.
Yes, that is the same drawing I have.  It's a 470 ohm (R88), not 47.
It is used as a voltage dropper because they are tapping off the +40V supply.  The other circuit has it's own winding and rectifier for a separate supply and has a 10R resistor.
I'm not arguing that the .1uf bypass cap may help, I just wanted to point out that the supply circuits are quite different and the 470R is not just for filtering.

Agree to try the .1 bypass as close to the DSP chip as you can and see if it helps!

Roly

Don't change the resistor.  As g1 points out it is also a voltage dropper for the DSP and should remain at that value.

I think that 8000uF is too big.  This will give a low frequency rolloff far lower than required, and may slow the rise of the supply to the DSP which may have implications for its internal power on reset function.

The fact that you remember the hum starting suddenly suggests to me that a power supply filtering electro on the DSP board has gone low/open.  One reason I have for suspecting this is that there was a period when makers of these small board-mounted electros were having some troubles with their electrolytes, and for some time a lot of these caps were turning up that were leaking onto the PCB.  I had a video projector brought in one time where the PCB was simply covered by goo that had leaked out of several of the caps, and I'm not known for throwing the towel in easily but this was clearly a lost cause.

I'm hoping that in this case simply putting back 20, 50 or 100uF across the supply will fix the problem.
If you say theory and practice don't agree you haven't applied enough theory.

domohawk

Copy on leaving the resistor alone.

To be clear the sudden hum was the loose J10. The hum now is back in it's "normal" state, however since I noticed this thing can really whisper with the dsp disconnected, trying to see if I can improve on the original "with dsp" circuit with some more filtering.

domohawk

Sorry for delay on results. Better late than never I suppose..

So building the caps into the J10 connected cable (by bridging power and gnd pins/wires) seemed to have little if any effect. However putting it directly across the diode bridge worked pretty well and cut out a lot of hum. I went with a 4100uF cap, since 8200uF seemed to have no added benefit, but was noticeably better than 100uF and even 1000uF.

I also wired in a switch in series with the J10 cable to simply break the power pin, and JB welded it to a cutout on the back panel. It's pretty dumb but works to completely kill DSP+associated noise (both hum and the slight digital "air" noise). Just make sure the amp is powered off when toggling this one or you get a great pop.

Thanks for all the help! I think I have the best of all reasonably possible worlds now with this amp.

Roly

Thank you for going to the trouble to close the circle on this one.  Too often people seem to have a successful repair, then don't bother telling us about it.

So from what you say it looks like the basic problem has always been in the actual supply - not enough filtering microfarads (and that would get worse with time).

Very well done, and thanks again.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.