HELP: OT for power transistors

[mac]

I can not get an OT nor data for a project that uses a single power transistor (BD237, TIP41 or similar) in class A mode, 2W-5W top.

I have some unused transformers and I need some advice to rewire one.  I rescued a little transformer with multiple inputs and outputs, all near 10 ohm. It works but given its small size it saturates a little.
My bet is that rewiring a 220vac - 12vdc wall wart having a secondary of 8 ohm and a primary having a ratio of 1:1 to 4:1 might work.
I'm in the right path?

ANy help is welcome. Ah, I do not have tools to measure, just my cheap chinese DDM and my ears.

mac

[Cliff Schecht]

Hmm.. The core material from a wall wart will work fine, I think your big task will be finding the right sized core. Why do you need a transformer on a bipolar transistor output? Your output impedance should not be that large. I'd like to see a schematic, maybe I or somebody else can help you more then.

[mac]

This is what I have on the breadboard,

home made darlington using two BD237, B=150,40, heat sink; small OT both primary and secondary 10 ohm, 1:1 turn ratio sounds better.
OT from 18V PS to collector, 3.9 ohm from emiter to ground not bypassed, base set to have ic=100-200 mA.
base bias, sounds familiar?

I tried another turn ratios, 2:1 and 3:1 but it saturates more than 1:1.
preamp is a fet clone of an epi valve junior. I posted about it recently.

As is, it is putting 1-2W into the speaker, and sounding much better than I expected. Even though the OT is crappy, it is adding some creamy tones.

Why do you need a transformer on a bipolar transistor output? Your output impedance should not be that large.

Just for fun!

mac

[brett]

OTs work well with BJTs. Don't know about JFETs.
There is magic in OTs.

[brett]

Oops! I should have read that more closely.  It's a Darlington, not a power FET.

First - a push-pull (class AB) amp will give a LOT more power for the same current.
If you want class A transformer distortion characteristics, you could use a class A driver transformer before the output section.  Such as a driver transformer as the phase splitter.  You'll see such a sheme in the famous "Deacy" amp.

Don't forget that you want the DC bias at the collector to be about 9V (1/2 supply).  You have Ic=150mA and Rc=10 ohms, so Vcollector is about 16.5V (18-1.5 V).  This is 6 to 8 volts too high, and barely usable.  You can raise the DC resistance of the primary to 50 to 75 ohms or raise the amount of current to about 1 amp so that the voltage drop across the primary is 8 to 10 volts. 

For class AB (near class B) push-pull amps I have used transformers with centre tapped 48 ohm primaries (ie 24+24:.  I think I used BD139s with 1 ohm emitter resistors.  Output power wasn't great, but the amp sounded sweet. 
good luck and have fun

PS I recently notice that Xicon have small transformers that are suitable as driver transformers and OTs for this style (and Deacey style) amps.  They are probably used in transistor radios and are available through Mouser, with parts No 42TU008-RC (1.5K:0.5k transformer for a class A driver section) and 42TU048-RC (48 ohm:8 ohm transformer for class AB or class A output).  The main disadvantage is that the rated output power is only 0.5 W.  Unfortunately, Mouser doesn't post to Australia, so I haven't been able to get my hands on them.

[Gus]

You might want to pull the lams and gap the transformer for the DC current that flows in it.  This is if I understand you post correctly, that you have the transformer in the collector leg of the circuit.

Is it something like the output section of this?
http://www.jlmaudio.com/neve_ba283.htm

[mac]

OTs work well with BJTs. Don't know about JFETs.
There is magic in OTs.

This is why I do not want to use a cap to couple the speaker.

Don't forget that you want the DC bias at the collector to be about 9V (1/2 supply).  You have Ic=150mA and Rc=10 ohms, so Vcollector is about 16.5V (18-1.5 V).  This is 6 to 8 volts too high, and barely usable.  You can raise the DC resistance of the primary to 50 to 75 ohms or raise the amount of current to about 1 amp so that the voltage drop across the primary is 8 to 10 volts.

Is it really neccessary to set bias at 9v? I ask because the Deacy does not seem to bias at 1/2 Vcc, nor the EL84 of my VJ.

Unfortunately, Mouser doesn't post to Australia, so I haven't been able to get my hands on them.

Every time I have to change tubes I have to place an international order. No good speakers, OT, PS, tubes, etc in my country...

You might want to pull the lams and gap the transformer for the DC current that flows in it.  This is if I understand you post correctly, that you have the transformer in the collector leg of the circuit.

Is it something like the output section of this?
http://www.jlmaudio.com/neve_ba283.htm

Yes, similar, except that I have a lot less parts: base bias resistors, 2 BDs, emiter resistor, OT.
An air gap is a good idea, 100mA is likely to saturate any average wall wart.

Gus, Brett, thanks.

mac

[davidallancole]

Whats your power supply voltage?  The voltage at your collector will be slightly less than the supply voltage due to the dc resistance of the transformer, which generally isn't much.  You don't need to bias to 9V if you are using 18V, it really just wouldn't work because of the low reistance of the primary at dc, the current would be prohibitivaly large.  I am not sure what the other guy was saying when he said to bias to 9V.

[brett]

 
oops.  Forget my stupid comment about bias.
The transformer primary is like a "spring" of impedance wrt AC, and Class A gain is about the impedance ratio (primary:emitter).

[Cliff Schecht]

Mac, you definitely want to get that transformers DC resistance higher. My recommendation is that you rebias the amplifier for 3/4 VCC in the collector (13.5 V) and use about a 50 ohm (primary) transformer. On the primary side, your 50 ohms of resistance gives you about 100 mA of collector current and a good amount of headroom to boot. Picking your transformers secondary winding is going to be the tricky part and is what I would muck with the most. I would try a 1:1 (50 Ohm to 50 Ohm) or a 5:1 (50 Ohm to 10 Ohm) first. I'm especially curious about the 5:1, although I suspect that it will drop amplitudes too low. You can also, of course, play around with the primary winding resistance. Anywhere from 25-50 Ohms is fair game IMO and would probably give a range of tones. Imagine if you could select between a "sweeter" sounding 50 Ohm primary and "grittier" 25 Ohm primary! Also, adjusting your emitter resistance using the two resistor/cap trick may also buy you some additional "features", that being an adjustable AC gain and bass boost. Make the bottom resistor (the one in parallel with the cap) something like 47 Ohms and place a 10 uF cap in parallel with it (giving a nice bass boost around 340 Hz) and make the other resistor adjustable from 0 to 100 Ohms.

[Gus]

I can't seem to find a book I own.  IIRC it was called the "Design of VMOS circuits with experiments"
http://www.bugbookcomputermuseum.com/Design-Vmos.html
IIRC in it is a SE mosfet 4 watt circuit a jfet and a power mosfet with a transformer in the drain leg.
I seem to remember it used a 24 ohm primary output transformer.

[mac]

Thanks for the replies.

I found a 220V: 48V (7 ohm) + 9V (1.2 ohm) transformer. I used the 48v as the primary and the 9 as the secondary. There is a little headroom but given the size of it it has more bass content, more inductance I bet. And it is sounding louder. There are still some nasty things but this is not an audio transformer. But the cream is there.
The 220v wiring can be used as a tone control by closing it with a small pF cap, cool!.

Considering your replies, and some measures I did on my epi VJ, 70 ohm: 8 ohm, I'd try a multiple primaries and secondaries transformer, for example, 8-50 ohm primary: 8-25-50 secondary.

Bypassing the emiter bias has a side fx, too much treble and some high oscillations, but this might be the breadboard.

ANother question, very diy. Anyone tried other materials in cores, like wood or dia/paramagnetics?

mac

[Cliff Schecht]

You could go for a lower loss toroidal core transformer. I can see a 5:1 impedance matching transformer making a lot of sense here. I'm glad I was right on the typical DC resistance of an output transformer being near 50 ohms, it makes perfect sense with the biasing conditions you specified.

[brett]

Hi
a toroid will saturate with a tiny amount of DC.  They are difficult to make work for push-pull amps.  For Class A amps an EI transformer ideally has an air gap.  You'll probably get away without a gap because of the low power.  Without a gap, I'd increase the size of the transformer (core area) by a factor of 2 or 3 to prevent saturation. 

A simple option is to use a power transformer rated for 2 or 3 times the output power you want.  Why does this work?  The inductance limits the low end response, and power transformers operate at 60 Hz in the US, which is fairly much the same as your cutoff frequency (low E is about 80 Hz).  Therefore, any PT works fine for AC all the way down to 60 Hz.  Multiplying your AC output mA by 3 to choose a PT rating allows for the fact that there is going to be as much or more DC current as AC current in the PT primary (and the power transformer is only rated for AC, of course).  Note that for push-pull systems, there is minimal net DC, you don't need a gap and you can use the full mA rating of the PT. 

The idea of cutting highs by adding capacitance is very cool.  Good work.
cheers

[mac]

This is a demo using the power transformer I described above, first secondary of about 7 ohm to transistor collector and to Vcc, the second secondary of near 1.2 ohm to my Laney HH 12" speaker; primary unused.

crunch demo

Brett, thanks for the info.

mac

[frank_p]

Nice tone.

[brett]

Using secondaries as primaries and secondaries is a killer idea.     

There is likely to be a compromise in selecting the VA rating of the PT/OT.
For low losses (due to magnetisation, eddy currents, etc) it would be best to use as low a VA rating as possible.  (e.g. for a 5W amp use a 3 or 4 VA transformer.)
However, the secondaries in such transformers are often made with fine, "angel hair" wire, and the DC resistance losses will be larger.

OTs usually have about 1 ohm of R for an 8 ohm tap.  3 ohms or more of resistance would be very inefficient for an 8 ohm speaker. 

first secondary of about 7 ohm to transistor collector and to Vcc, the second secondary of near 1.2 ohm to my Laney HH 12" speaker; primary unused.

Those resistances are near perfect.

Maybe we can measure the resistance of some PTs and find those with appropriate VAs and low R ?

Just make sure that
1. the secondaries are isolated from each other (or their connection can be cut at a terminal), and
2 the voltage ratio is somewhere around 1.5:1 to 3:1 (impedance ratio of 2:1 to 9:1).  For low voltage amps, lower ratios are good because they mean less voltage (and power) loss.
cheers

[brett]

Hi again
I'm home now and have made some measurements on PTs.  It seems that several I have with VA ratings of 5 to 15VA have really good, low resistances.  (It's a pity they mostly have "flying leads", where you can't cut and separate the secondary at the terminals.  Damn!)

Here's a thread and photo essay for anyone thinking of winding your own low voltage-low impedance OT.  http://www.diystompboxes.com/smfforum/index.php?topic=53786.0 

cheers.

[mac]

The size of the core is, 7 x 6 x 2 cm. Input 220vac, output 48vac and 9vac, separated.
I forgot to mentioned that I found it almost buried in an empty lot, it is rusty.

For this small amp my bet is that any transformer of this size and voltages can work.

Brett,
I have been searching for this link for months! I read it some ago and forgot to bookmark it. Thanks.

Using secondaries as primaries and secondaries is a killer idea.      

And the primary can be used as a tone control by closing it with a small resistor and/or cap jeje.

mac