Matching DIY 13003 Darlingtons

[mac]

I have a bunch of 13003 and 13003D (with C-E diode), TO-126, 13<hfe<20.
I want to use them in a 18v push pull transformer coupled amp, circuit adapted from my old Crown CRC-530 radio recorder which runs at 7.1v.

The idea is to make two Darlingtons using a 13003 at Q1 and a 13003D at Q2, and B-E resistors.

I understand that this kind of circuits call for matched Vbe, hfe output pair... although the Deacy has unmatched ACs...

My question is,
Is a Darlington of hfe (20,20) similar to a pair (25,16) if total Vbe are equal?

mac

[PRR]

Does the Crown(*) use Darlingtons in the power stage? Most early transformer-coupled amplifiers did not
(*) US lurkers: this is not the same Crown as the famous DC-300 amplifier. Brand: Crown Radio Corp.; Tokyo alternative name Asahi Radio Mfg. Co., Ltd

[Rob Strand]

Does the Crown(*) use Darlingtons in the power stage? Most early transformer-coupled amplifiers did not
(*) US lurkers: this is not the same Crown as the famous DC-300 amplifier. Brand: Crown Radio Corp.; Tokyo alternative name Asahi Radio Mfg. Co., Ltd

Something screw up with my post.

I was going to say the same thing.

I couldn't find a schematic. I found a CRC-5800 or something and it used single transistors.

The gain of Darlingtons is strongly dependent of current and the characteristics of *both* transistors.   It's creates a lot of headaches, what current are you matching at?!

If you were to use Darlingtons you would need a very low BE resistor of Q2 so the bias was set by Q1 alone.     If you need more output power then the BE resistor would have to be quite low and even then it's just passing the power dissipation from Q1 to Q2, not gaining power from two transistors.  What you really want is parallel transistors with large emitter resistors.  Fortunately a lot of those designs use large emitter resistors so it might be possible (with some care regarding matching).   Nonetheless it's not widely used form and maybe for good reason.

[mac]

Does the Crown(*) use Darlingtons in the power stage? Most early transformer-coupled amplifiers did not

My Made in Japan Crown has D355 silicons, extended to-92 case. 2N3904 or BC237 on steroids. No heatsink.
Idle current at the shared 1 ohm resistor is 15.5ma. Running at 7.1V that's 50mW each.
At 18V it needs more idle current to bias properly, +25ma each giving half a watt or so.

It's creates a lot of headaches

I have BD, TIP and ADs, but I also have these unused very low hfe 13003 that combined give a gain of 300-400.
Let's say I want to try them just for fun.

One can get the same total hfe with two similar pairs, say (20,20) and (21,19), or very different pairs, say (20,20) and (80,5).
Just curious of how much aspirin do I need in the second case? 

Asahi Radio Mfg. Co., Ltd

Cool. Thanks!

mac

[PRR]

Cool.

Some history on their site.

http://www.asahikeiki.co.jp/en/profile/history/

• 1950      Fuji Keiki Co., Ltd. stopped its business, and Asahi Keiki Co., Ltd. was founded
• 1958    Developed and manufactured transistor radios
• Developed prototypes of TVs and microwave ovens
• 1960    Started to operate the Fukushima Plant (radios and thermostats)
• 1963    Fukushima Plant started up as Fuji Onkyo Co., Ltd. (manufacturing radios integrally)
• 1971    Started to sell digital panelmeters and digital testers

They sold a LOT of digital panel meters, but today they do a lot of thermostats.

[Rob Strand]

I have BD, TIP and ADs, but I also have these unused very low hfe 13003 that combined give a gain of 300-400.
Let's say I want to try them just for fun.

One can get the same total hfe with two similar pairs, say (20,20) and (21,19), or very different pairs, say (20,20) and (80,5).
Just curious of how much aspirin do I need in the second case? 

I'm all for using old parts.  I've been doing the same lately.   Trying to get old crap to run at 100MHz by being clever.  You need to work hard for your money compared to buying modern parts.

I don't think the second case will work.   What you need to do is build something and check the bias currents.   I suspect the first run through isn't going to work so great unless you have large emitter resistors even with the well matched devices.

I was very interested to find this,
https://sound-au.com/articles/pwr-amp-dev.htm

See figure 2.2 and figure 2.3.  Surprisingly for figure 2.3 Rod Elliot pretty much confirms what I said earlier.   He got his design to work with a separate bias pot for each half.  Also notice he used some thermistors to prevent thermal runaway - with large emitter resistors you might get away without using thermistors.  So you can get it to work with some pushing and shoving.

[mac]

I've tried the 13003/13003D Darlingtons.
Two pairs, (hfe1, hfe2) --> (19,13) and (18,15). Both having a final gain of about 215 at 5ma-10ma, and equal Vbe. While measuring I added a 220 across the B-E of the second transistor and gain dropped to almost the value of Q1. I decided to not use it for now.

It didn't work  ... until I realized I had a single diode at the bias network
I removed the diode.

It works as good as a pair of TIP41 hfe:150 I was using. I'd say it's a bit hotter.
I have some 2N2369 I'm going to try at Q1 to have a hfe:300-400 pair.
I guess I'm 70 years late... 

separate bias pot for each half.

I always start with a shared resistor and after a while I check C to C voltage. If there is a difference, ie an appreciable net current through the transformer, then I use a 10 ohm trimpot to set C to C near zero volts.

In my experience the shared resistor works in most cases even if transistors are a bit unmatched.
Equal resistors for each emitter are better.
It's the driver transformer which gives me headaches most of the time. I'll be asking about drivers soon.

See figure 2.2 and figure 2.3

Who I am to disagree with Mr Elliott but my Noblex Carina TN2-E adapted for guitar and running at 12v (twice the original voltage) is way better and creamier than any of those ultra low distortion, high power, state-of-the-art, HiFi SS amps 
Seriously, with a Rangemaster or TS in front the hot Toshibas 2SB56s breaks a 12" nicely.


I'm a big fan of those old radios and recorders. Maybe it's because my first amp at the age of 6 was a Crown CRC-530 which I used to drive harder with a Pioneer stereo. Very Deacy, and still working.

mac

[mac]

A picture of the Thing using darlingtons and fets Toshiba 2SK246 and 2SK117 @ Q1,2
Best bias for the Driver Transformers is 0 ohm at the emitter! Have to figure out why. The pict shows a 10 ohm used for feedback.

Transformers are from a Crown tape recorder, same as those in my radio recorder.

PS is a rectified transformer, 15VDC.
I avoid voltage regulators, they introduce strange "artifacts", maybe a breadboard thing.



It sounds great. A treble booster or TS at the front are a must.

mac

[R.G.]

Darlington connections of discrete devices can be funny. At low currents, simpler-geometry bipolars exhibit an increasing gain with increasing collector currents. That's the reason that transistor data sheets sometimes list hfe at several currents. It's also the principle behind modifying the emitter current of a differential pair to make a multiplier stage.

That means that using the same part-number discretes as both first and second transistor in a Darlington connection can give significantly different gain that expected by multiplying the individual gains together - the first stage operates at a current of  1/hfe of the second, so its gain may be much lower tan expected.

Diffusion and geometry can be jiggled to affect this in integrated Darlingtons and in making high gain power devices that don't have as much gain falloff at low and high currents. But for discretes, you may find low total effective gains from using the same device type for both first and second transistors. Of course, if you're getting enough out of the stuff you're using... well, never mind. 8-)