Menu

Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Messages - Roly

#2101
There is a trick using SCR's that can be used to reduce the voltage of a transformer for a power supply (see attached).

The gate trigger voltage is obtained from a divider between the transformer peak voltage and the output voltage.  When the output voltage rises there is a point where the voltage on the gates doesn't rise high enough to trigger the SCR any more, thus limiting the output voltage.

It has a few "gotchas" however;

- at higher voltages the gates may need to be protected by diodes

- the output voltage is a function of the gate trigger voltage which is itself a function of the SCR temperature, so this simple circuit may show a high temperature sensitivity

- SCR's and audio generally don't make for good partners and there may be considerable, and objectionable, high frequency buzz in the output
#2102
Tubes and Hybrids / Re: Rectify 110v?
June 08, 2012, 08:42:29 AM
Please forgive me folks, but every time I see the title of this thread it gives me a bad dose of the jeebies.

Rectify the mains?  NO!  DO NOT do this!   :duh

There are situations where high voltage DC supplies are derived directly from the mains without an isolation transformer (mainly in industry), but guitar amps are not one of them.

The reason is that the strings of a guitar are electrically connected back to the amps chassis, and with direct mains rectification it is not possible to be sure the chassis isn't connected to the active side of the mains, with almost certainly fatal results for the guitarist.   xP

Another way of getting high voltage DC that is isolated from the mains is to use a voltage multiplier.  Here is a discussion of such multipliers specifically for use in guitar and other audio applications;

http://www.ozvalveamps.org/ava100/ava100psu.htm
#2103
{As JM says, plus...}

The resistance at the drive end is generally more important than the resistance at the pickup end.

Generally with spring reverbs, the longer the spring, the better; the more springs, the better.  Many console organs are being junked these days that have both long and complex spring reverb tanks, and also rotary speakers which can make guitar sound way cool.
#2104
Nope.  Try this.  See anything you recognise?
#2105
"30 minutes" strongly suggest that this is basically a thermal problem.  A quick Google of this amp produced several people apparently having thermal/cooling problems with this one, including a post here back in 2010;

http://www.ssguitar.com/index.php?topic=1513.0

...and one on MTS where his amp caught fire!   :o

Quote from: http://acapella.harmony-central.com/showthread.php?2413495-Randall-G3-Rh300-vs-Rh300
a host of problems related to basically sloppy engineering
...
oh and also turning the f**k off when you turn it up to a certain point. That was the real kicker for me when I had mine. Get loud? NOPE, turns the f**k off.
-----
a friend of mine bought the plus model due to the bad reputation of the non-plus models. it sounds great but it has the same problem of turning off once the volume goes past 4.
-----
a couple weeks ago and I played two G3 in a row, factory fresh, with the "turn off" problem,

The output stage appears to be MSOFET-based.  From the above it seems that this model amp, at least, has some sort of generic problem, most likely related to the cooling of the MOSFET output stage.  Unless somebody turns up a known fix (such as added cooling) I don't think there is much that can be done with what seems to be basically bad design problems - sorry.
#2106
{Still do as Phil suggested as best you can, your own way.  One tip I can suggest is that you make a drawing/tracing of the copper side area of interest, then try to locate the components on the upper side onto that drawings.  Done as carefully and accurately as possible, that alone would be very helpful.  When trying to locate components a powerful desk lamp shining from the copper side can give you an "X-ray" view.

Resistor colour code charts;
http://www.google.com.au/search?q=resistor+colour+code

On-line calcs;
http://www.csgnetwork.com/resistcolcalc.html
http://www.hobby-hour.com/electronics/resistorcalculator.php
}

Okay, a few things you can check/try before posting photos.

See bottom view, attached.  Measure the resistance between ground "Gnd" and point "A" while varying the gain control.  The resistance to ground at this point should vary between nothing and whatever the value of the gain pot is (a few k ohms?).  If this appears to be open circuit try remaking the solder joint to the pot directly above "A".

Resistance swings up and down?  Good.

Next do the same thing from ground to point "B".  In this case the resistance should go from the upper value found above, plus about 1k, to around 820 ohms.

Don't get this result?  Then the 820 ohms (grey/red/brown) resistor is faulty, (but this is quite unlikely).

Do get this result?

Try replacing the electrolytic cap (shown in green circle in the top view) with the same value (this doesn't have to be exact, 10-20% more capacitance and/or voltage is fine).  Note value and post here as soon as you know.  This cap is polarised and must be replaced in the same polarity (use marking pen to mark on the board where black negative mark is on the cap before you remove it).

Cured fault?  Post and we can all enjoy a beer.

Same as before?

Now we get out of repair territory into mod territory.

When you go to buy the new cap, also buy the following resistors (if you don't have to hand or can't scrounge from some dead radio or such); 680 ohms, 560, 470, 390, 330.

Replace the 820 ohm resistor (blue circle in top view) firstly with the 470 ohm (don't cut the leads short, mount it up in the air a little bit, say half and inch).

The gain control should now produce more gain at the top end.

You may get the following results; just right, not enough, too much, wild oscillations near the top.

If not enough, try a lower value, if too much try a higher value, if wild oscillations try a much higher value.

Once (if) you find a value that gives you satisfactory results you can remount in down close to the board like the original (and make a note on the board of your mod, original value, the date, and your initials).

Post results.  {Drink well earned beer.  :dbtu: }
#2107
Paolo, is there anyone who knows something about electronics you could get to give you a hand tracing the input circuit, science teacher, local radio ham, hackerspace etc?

All we really need is an idea of the circuit from the input sockets past the gain control to the main volume control.

Failing that we need two photographs of the input end of the board which need to be;

- taken with a camera that has macro "tulip" mode, important to get them crisp and in focus

- square on, directly above the middle of the board between the "gain" and "volume" controls, not at an angle

- fill the frame with the end of the board (see example)

- no flash, and lit from two or three different directions so that shadows are minimised (e.g. three disk lamps close left, right and in front, positioned so there is no reflected shine)

- both sides of the board like this

- do not reduce these, post full size

Also need two shots at different angles of the backs of the input sockets that show how they are wired (as above, but these can be reduced).


Meanwhile; making heavy use of my crystal ball, here is a bit of what I think the input and part of the first preamp looks like.

If there isn't a dry joint in the gain setting path, then the most obvious suspect has to be the can electro (but I still remain to be convinced that it isn't working as designed - unsatisfactory maybe, but as originally built).
#2108
Yeah, I have no strong opinion here except that something doesn't fit, and I've always found it profitable to follow the bit that doesn't fit.

E.g. to take JM's thought that the real fault here may actually be a cracked track, not a boofed FET, and that this has only been incidentally "fixed" in the process of replacing Q2.

On one hand the symptoms as reported pointed to Q2; but now, on the other hand, we have replacing Q1 as an effective repair, and I'm at a loss to see how the initial reported fault of "stuck in crunch" could be caused by a fault at Q1 if the switching rail at the join of the 1Meg and two diodes was actually going to -6.5V.  ???

I suspect there is another boot yet to hit the floor here.   :-\
#2109
Amplifier Discussion / Re: OCL power amp stages
June 05, 2012, 04:55:38 PM
{Since this circuit has been posted...}

Actually it's fairly similar to my Twin-50 which uses a 25-0-25VAC tranny to give rails of +/-35V and just on 50 Watts into 8 ohms.

I had to laugh when I saw the "150W" title because anybody who thinks they are going to get 150 watts out of a pair of 2N3055's is fooling themselves.   :duh

The supply is shown as +/-42V which is fine as long as you don't actually drive the amp because at full swing each output transistor has to withstand 2 * 42 = 84 volts, while the Vcer rating is only 70V max - booffo!.

The speaker impedance isn't specified but looks like 4 ohms was intended, and again 2N3055's aren't happy with loads below 8 ohms because the current rises to a point where the Hfe drops quite seriously.

This is sadly typical of many circuits that are published on the web that somebody has drawn up but never actually built, leaving it to trusting readers to suffer the bruising experience.   xP

All the other transistors could be substituted by BD139/BD140's.

D4 and D5 should be thermally coupled to the output transistor heatsink (which even for 50W into 8 ohms needs to be quite generous).

Ra can be a 47 ohm 2W resistor and L can be as many turns of 2mm enamel covered wire as will fit on the body of Ra (about ten), or both can be replaced by a 0.22 to 0.5 ohm 5W wirewound resistor.

Rb can be 10 ohms 1W and C a 0.47uF 100V greencap.  Note that this should actually be connected to the join of R15 and R16, not as shown.
#2110
Nope. Obscure-tronics; all I get from Google is Kermit and Miss Piggy.

Quote from: Paolo
It was like this when i bought it off ebay a few years ago.

It is quite possible that it isn't actually faulty, but was built like that, which is why finding an actual circuit would be very helpful.  It is pretty unusual for an amp to simply lose gain and still sound clean, without distortion or crackles &c.

If you could trace the circuit from the input, past the gain control, and up to the main volume control, and post that here it would be a great help.

You can also try two other approaches, signal tracing, and signal injecting.

In the first you apply a signal to the input, then use a lead connected to the input of another amp (via a dc blocking capacitance) to follow the signal and see how the gain ramps up  (or doesn't) as you move from input to master volume.

Signal injection is similar except that you inject a signal starting at the input and moving to the volume control.

But in both methods the idea is to get an idea of how much gain each stage is providing, and if one stage isn't providing any.

A possible single fault is that an emitter bypass capacitor has gone open for some reason, which would give a clean reduction in gain.
#2111
Quite so, but I've been pondering the circuit trying to rationalise a couple of things that don't gell here.

In post #4 OP says "The voltage immediately after the 1mrg resistor changes from -6.5v to +6.5."

Now if that is the case how could the crunch channel be on?  The original complaint was that it was stuck in crunch, post #1 "The crunch channel is always on.", i.e. Q2 wasn't turning off.  If this was because a failure of Q1 was preventing the control line going, say, more than a diode drop or two -ve, how do we have -6.5V after the 1 meg resistor?

If Q1 was shorted the amp should have been switching between crunch and nothing.
#2112
"No real info" is right - I can't see any trace of the maker never mind the model, but as Phil says it won't be rocket surgery, and failing an actual circuit some good clear pix inside around the gain control area and where that connects to the PCB might help.

IrfanView - used it for years now prepping pix for my website and I couldn't live without it.   :dbtu:

When it comes to internal pix for diagnosis, if they are well lit and in sharp focus it is generally better not to reduce them as detail is lost.
#2113
Are you sure you mean Q1, the one across the clean channel, and not Q2, the one in series with the crunch channel?
#2114
Please sketch, scan and post your current power supply arrangement.  I can think of at least a couple of quite different arrangements that would fit your description.

The most obvious first thought has to be buy/scrounge/swap a more suitable transformer.

#2115
Good one Joe; I love it when a chip (like a regulator) gets used for something the maker never imagined.   :dbtu: