Crate GFX 30 solid state amp problem

[betweenthees]

I have a Crate GFX 30 solid state amp I'm having a problem with. After it is on for like a min it makes this white noise.(like the static sound of a "snow" tv channel). You can play through the amp before and after this happens. When you turn the amp off and then back on right away, the noise stops, waits another min, then starts again.

I cant find a schematic for this amp. (would realy appreciate having one)
I can post any other info you may need.
some basic onfo on the amp
http://www.crateamps.com/pdf/manuals/GFX30_OM.pdf

Thanks for any input

[betweenthees]

This post is going to lack some proper terms and many saftey features but....

i got out my multi meter and began testing everything i could get my tips to touch. There were 5 Jfet transistors in the area that i thought was the preamp. all but one gave similar readings with the amp on and off. The last FET showed the following...

amp turns on- V= around 20v
this slowly increases to around 42v and the "noise" begins
so i placed a 10Mohm resistor between the 2 lugs that had the increasing voltage.
this stabalized the voltage to around 36V (below what seemed to be the cutoff for noise)
What im trying to do now if figure out what realy is wrong. what i tried to explain above was just me throwing ideas all over myself looking for clues.
My first move is going to be to remove the FET and replace with a new one- then work from there.

[phatt]

Sounds more like a psu problem not a signal transistor issue.

Look for how/where that voltage is derived,, ei some just use a Zener to clamp the voltage. if that goes futt the voltage will climb to the next highest voltage in the psu.
Look for a drop resistor (often a large wattage type,, ie 2W or even 5Watt next to a diode)
Others here might know more about your particular amp,,, be patient. :tu:
Phil.

[J M Fahey]

Fully agree with Phatt and add: I think fets are used as switches there, not as amplifiers.

all but one gave similar readings with the amp on and off.

"Similar readings" isn't a very clear statemente, isn't it?
You mean voltage readings , resistance readings, what?
1)post the Fet codes
2)call legs 1,2,3 (as in datasheets), seen from below, flat side up, left to right
3)measure and post voltage on each leg, amp on, since they switch something check whether pushing switches (panel or foot) changes any voltage.
Post them in some kind of table; example:
Q1, 2N5457 , Leg: 1/2/3 : 0V/0V/-15V (Clean ch. on)
Q1, 2N5457 , Leg: 1/2/3 : 0V/0V/0V (Clean ch.off)
and so on.
Post the "strange" one as:
Q1, 2N5457 , Leg: 1/2/3 : 0V/0V/-30V (Clean ch. on) The -30 rise to -42 in 1 minute, hisses badly over -39V
4) I'm pretty sure your Fet is fine, but being probably a 25 or 30V part, becomes a very noisy "Zener" above that.
As Phatt said, the problem must lie in the abnormal voltage is hitting it.
Get used to repairing amps without schematic  :grr because that's one of Life's hard realities.

[betweenthees]

Thanks for the replys- they got me thinking more clearly about what i was looking for

1) The transistor I was speaking of before was a j175
datasheet-- http://www.datasheetcatalog.org/datasheet/philips/J174_175_176_177_CNV_2.pdf

I was taking voltage measurements from the "gate"
the V would increase from 0-42v < 1 min after turning on

2)took your advise and traced from the trans. to the source and found a resistor and a diode.
4M ohm resistor and a 1n4 148 diode (in parrallel)
These had a constant 46V going in with (increaseing) 0-42v < 1min going out

I replaced the diode and resistor with new parts (didnt help or hurt anything)

3) lm317 voltage regulator--
http://www.datasheetcatalog.org/datasheet2/1/03cgthpfat4t4ly5kfp5lpwladfy.pdf
I happend to have one of these laying around so I placed this into the curcuit before the diode resistor set. (so that the output of the lm317 would go to the input of the resistor and diode--> that would go to the transistor gate)

I set the regulator to 40v (using the formulas in the datasheet to choose R1 and R2 i think I used r1=240 and r2=6.8k ohms) I feel as if i jumped over the real problem, but I have the Voltage ceiling I want where i want it (remember 41v+ was overloading the transistor and causing terrible noise)

I tested for a while and found all the controls work fine. The regulator works and never allows more then 40v to pass. The output V of the diode and resistor maxes out around 39V and that goes to the Gate of the transistor.

4- i Emailed Crate and asked for schematics- It seems like i can get one from them if i fill out the correst waiviers. It may take some time, but would anyone like a copy of them if i can get it?

[J M Fahey]

Hi betweenthees.
Now things are clearer, and without schematic.
1)As you see on the datasheet, the J174 is a 30V part (as I guessed), specifically: Vgs=30V absolute maximum.
Of course parts often have some extra tolerance, some might even stand those 42V, but you can't trust that at all.
Overvoltage breaks the junction barrier, not burning it because it's current limited by a resistor, but becoming an excellent source of white noise.
From what you say, that Fet is an On delay mute, probably grounding the power amp input for a few seconds, until its gate gets positively biased (that's why they use a P Fet there)
Your regulator prevents overvoltage there.
This was taken from a GX65 schematic and must be about the same you have.
As you see, a high value resistor (over 1 megohm) plus a diode go to +B, replacing them changed nothing, so they were not the problem.
Also it takes 1 minute to reach +42V, which means the capacitor is fine too; that leaves R52 as tyhe main suspect. It must be open or the tracks to it may be cracked somewhere or the solder itself cracked.
One end must go to the Fet gate, the other to ground.
The voltage divider action (2M2/680K) should supply around +10V to the gate; it being open will slam it with the full +B, causing heavy noise.
The values and part numbers will not match in your amp, but the general idea is that.
Probably you can pull the regulator now, by repairing the original fault, yet I must congratulate you on finding a working solution on your own. :tu: