Randall RG100ES Question

[MRMcGovern]

I've listed out all of the parts that I need to build an RG100ES from scratch, but there is one component I'm not sure of. I believe it's a thermistor, but no values are labeled next to it. Can someone provide some insight?

(Looks like a squiggly line with a circle around it)(All the way on the right to the left of the two diodes in the center)

http://el34world.com/charts/Schematics/files/randall/Randall%20RG100.PNG

Please help! Thanks!

[Loudthud]

The device you are wondering about is in parallel with D7. It is probably a thermister, a temperature dependant resistor. It is included to regulate the bias current in the output transistors as the heat sink temperature increases.

One of the challenges in designing a power amp like this is keeping the thing from self destructing. The bias current in the power transistors will tend to increase as they warm up in a vicious circle until they get too hot and fail. Some means of thermal feedback is required that will maintain or slightly decrease the bias current as the power transistors warm up. The 1N914 diodes might be sufficient if they can be mounted on the heatsink. I like to use a TO-92 transistor held face down on the heatsink with some kind of metal clamp.

In practice you build the amp and get it working. Set the bias when the amp is cold, run the temperature up by running the amp into a dummy load until the heatsink get nice and hot and measure the bias current again. If the bias current is within reason, the design is done.

[Roly]

...and...

This particular circuit has quite a bit of compensation/adjustment.

The resistor-in-a-circle marked "TRM" is a thermistor and a bit of a minor problem because it doesn't specify what thermistor, and there are a wide range available, HOWEVER strictly speaking this might be gilding the lily because few amps actually use a thermistor, so there are ways around needing it.

There is also the 250 ohm "Trim" bias preset pot, and even R51 "47 to 100 ohms" which is a select on test in the factory, plus the 22 ohm in parallel with one of the 1N914's, so there is a lot of adjustment available.

What I do is forget about the diodes and use the Base-Emitter diodes of a couple of transistors, BD139's being my favorite.  These are "thermopad" style with a hole through the middle which makes mounting them a whole lot easier.  These two are then mounted on the heatsink as close as you can get them to Q15 and Q16 (being careful to keep their metal pads insulated from the heatsink).

This way the temperature compensation diodes are closely coupled to the things they are supposed to be actually sensing, and you should be able to just forget about the unknown thermistor.  The transistor Base is the diode anode, and the Emitter is the diode cathode.

On initial power up the trim resistor should be set for minimum resistance and you should use a limiting lamp for safety.  If all is well you should have little or no idle current (voltage drop across R57 or R60).  You then advance the trimmer until you get the required standing bias current, generally about 50mA;

E = I * R

0.05 * 0.27 = 0.01 or 10 mV across R57 or R60.

From here you can find three cases, under-compensated, over-compensated, or just right.

Under compensated the idle current will rise as the heatsink temperature rises.  This is the most dangerous situation because if it gets hot enough you might get thermal runaway and a cooked output stage.  With two diodes/transistors on the heatsink this is pretty unlikely, what is more likely is...

Over-compensated, where the idle current falls as the temperature goes up,  This will cause crossover distortion when the amp is hot.  It won't hurt anything but it won't sound too good either.

Just right compensation is where the idle current remains pretty stable whatever the temperature of the heatsink, say within +/-20%, and this is what you need to aim for.

If it does turn out to be under-compensated you can add a third diode/transistor in series and reset the trim.

If it turns out to be over-compensated you can either put the 22 ohm resistor back in, but start with a higher value like 220 ohms and work down, or you can try taking one of the diode/transistors off the heatsink and mounting it back with the general circuitry.

This whole thing is a bit of a chore because each change you make you need to take the heatsink from cold up to hot and back, but it's well worth taking the time to get it right because the amp will sound better overall for it, and you only need to do it once.

HTH

[MRMcGovern]

You guys are awesome, thank you so much for all of that information. I'm very new to this so a lot of what you said is way over my head as far as details go. I'm sure with a bit of studying I can start to understand the details.

If I may ask a dumb question. What components need to be mounted to the heatsink? Sounds like diodes and transistors from the power amp? From what I understand, as long as those components remain cool enough I may not need the thermistor?

[Roly]

If I may ask a dumb question. What components need to be mounted to the heatsink? Sounds like diodes and transistors from the power amp? From what I understand, as long as those components remain cool enough I may not need the thermistor?

Well we are here to walk you through it, and to help if you strike any difficulties.

I once remarked to an (PhD) electronics mate that I was a bit of a pessimist.  "Pessimist?  Anyone who starts out with a bag of parts expecting to end up with an amplifier has to be an optimist!"

Well you've got a couple of hundred years of amp experience on tap here - make good use of it and open a "build thread" where you record your progress, posts and pix.

Firstly people will offer hints and tips on construction as you go.

Secondly we will all have a record to review when it comes to powering it up for the first time (it can be a bit like launching a ship or rocket or something, there's a right way to do it without smoke or nasty surprises;

e.g. see "Bringing the amp up" down http://www.ozvalveamps.org/ava100/ava101lamington.htm



None of us were born knowing this stuff; the only "dumb" question is the one you don't ask.  I have had lots of trainees over the years and one thing I've observed is that there is something somewhere that each doesn't "get", but that it's always a different thing and can be  anything and ranges across everything.

Last night I got asked to do some homework on a series/parallel resistor network, and the first step of reducing a pair of parallel resistors to the equivalent single resistor was proving a hard concept to grasp.

Consider;

1. Constructing an amp from a kit with full instructions, like a Heathkit.

Pretty easy and requires no understanding and minimal constructional skills.

2. Building an amp from a circuit alone.

More difficult but still requiring little understanding but more skills.

3. Building an amp of your own design.

Fairly difficult, requiring some understanding of electronics theory and fairly high level of mechanical and mental skills.

4. Understanding how circuits work.

Range of difficulty, low to high, reasonable mental skills required.


When you hit level 4 you are giving birth to new ideas such as the Vari-GronkRC or the Lamington, however you can build three valve amps and a dozen stomp boxes and never hit level 3 there is no "yagotta".

But if you want to understand then there is some yagotta; and these days we've got many good sites and Wikiwatsit and Google.  {boy do I wish I had access to today's Web, sites/forums, datasheets, circuits, theory, etc., when I was building my first amps in the 60's!}


Anyway, buildthreadbuildthreadbuildthreadbuildthread...  :dbtu:


---

If it's a semiconductor and it gets hot, it needs a heatsink.

The main output power transistors Q15 and Q16 for certain.

Their drivers Q13 and Q14, most likely.

The pre-pre drivers Q11 and Q12, mmmmmmm maybe with the others on a big well-placed heatsink, maybe just individual "flags", or maybe not needed.

The Voltage Amplifier Stage (VAS) Q10 generally sits on top (or below) the thermal diode compensation string, possible flag, see how it goes.

May I suggest that for reference the actual diodes are D7 and D8, and the post mod thermal transistors-as-diodes be QD7 and QD8?

We have the "A" +ve supply rail, and the "B" -ve supply rail, and in between, connected to the speaker we have the "half-rail", so called because in correct idle operation this sits within ~+/-200mV or the nominal ground point (minimal DC trough the speaker, the amplifier has "re-balanced" to zero output).


---

I may be wrong about this ( +/- where you are in the world), but I rate identifying and obtaining just the right thermistor for this circuit as practically zero.

Changing the thermal compensation network is, at this point, the only real option you have; this bit of the circuit is going to have to be a bit redesigned to avoid the thermistor, and possibly incidentally improving its thermal compensation over the original.

---

A bit of background: if we start at the speaker and come back along the half-rail and go up R57, the emitter resistor of the upper output, we find a voltage across R57, then a diode drop of about 6-700mV depending on temperature through Q15.

From the Base of Q15 you go up the EB of Q13 and gain another 6-700mV diode drop, and again at Q11, so now we are 3x 6-700mV above the half rail, with a tempco of 3x 20mV/C.

The other way, downwards from the half-rail, it's only Q12.

But that makes the bias voltage just to bring it all to the threshold of conduction a thermally-compensated four diode drops.

And that's what we have to arrange.

The main bridge rectifier may also need some heatsinking consideration (so try and find a bridge with a hole in the middle like a PA40).

[Loudthud]

I would be beneficial to see a photo of the Randall amp guts. This will give a general idea of the size of the heatsink and what parts are mounted on it. It would also help to find out how Randall suggests the bias be set. Q15 and Q16 operate in class C (this is unusual) and only turn on when the load demands more than about 1 Amp. Q13 and Q14 are class AB outputs that probably have a bias current of 50mA or less, more than likely 10mA.

Another odd thing about this design is the high source impedance of the output. Voltage feedback only sets the quiescent DC output voltage but signal feedback comes from output current sensed by R63. The stability of this design is unknown.

[Roly]

Been pondering the RG100ES preamp circuit.

Would build nicely as modular boards for each section, mounted on a rail so they can be flipped up in circuit for testing/mods.

At the Fx Return socket, instead of using the back contact to Loop it, use a good quality mini-switch right next to it with suitable markings.

Q15 and Q16 operate in class C (this is unusual) and only turn on when the load demands more than about 1 Amp. Q13 and Q14 are class AB outputs that probably have a bias current of 50mA or less, more than likely 10mA.

Thank you for picking that up,  :dbtu:  I missed it this time around  , that's quite right, in this amp the final transistors act as "current dumpers" and only operate at higher signal level, the low level coming from the "drivers".  Quad Hi-Fi amps used something like this back in the 80's.

The bias is set to suit the "drivers", and the "output" dumpers just follow along as needs be.

{sorry}

[J M Fahey]

I have gut pictures somewhere but the main point is, it has a horizontal "Twin type" chassis made out of thick aluminum (around 2 mm) , the 4 TO3 transistors are directly bolted to it.
Although horizontal, the huge surface and free air movement make it a very effective heatsink.
I also had doubts about a pure current drive, meaning the amp goes open loop if no speaker present.
But these amps were very well built (I mean the original American ones) and seem to work for decades, so the design is time tested as sound.

[JHow]

I have a 1983 Randall rg80/100 es.   I can take some shots tonight if that helps.  The heat sink is as JMF described, the whole chassis itself.  It's very thick aluminum.


Sent from my iPhone using Tapatalk

[Loudthud]

I have a 1983 Randall rg80/100 es.   I can take some shots tonight if that helps.  The heat sink is as JMF described, the whole chassis itself.  It's very thick aluminum.

What we need is a shot of the PCB that shows what is mounted on the chassis, the driver transistors, diodes and/or thermistor. I looked around on ebay but all I could see was that the TO-3s were mounted on the chassis. I found the operator's manual. It says the output damping factor is 2 to 4.

Link: http://el34world.com/charts/Schematics/files/randall/Randall-RG80_100%20guit%20amp.pdf

[JHow]

Okay, so here is what mine looks like.  YMMV.  The thermistor is on the chassis, between the four outputs.  The 914 diodes are by the connectors and the tip31c in shot #2, the 1n4003 diodes are near the emitter resistors and between the connectors.  the thermistor is red and white and has a marking 750A703-2.

I realize I won't win any awards for photography, so let me know if you have questions about what you see.

Hope that helps.

[JHow]

Here is another of the board...

[JHow]

thermistor..

Added: The thermistor is mounted to the chassis with a bolt with star washer and the leads go to a tag strip and then wires back to the board.

Also, as loudthud mentioned above about the rectifier, it is not shown here but is also mounted to the chassis as he suggested.

The chassis is at least 1/8" thick.  There is nothing flimsy on these amps.  The circuit board is thick, too.

[Roly]

Notice that on the circuit the thermistor is right next to the diode, but in reality it is some distance away - through a connector.  And what would happen without a diode on the board if the connector got unplugged?

One well-boofed output stage.   

[J M Fahey]

VERY good, VERY well made amplifier.

Not surprisingly, it was a product by Don Randall, Leo Fender's partner and sales/distribution manager since forever.

He was clearly well versed in good quality and manufacturing control.