Joe Rasmussen "Trans-Amp" - 40 Watt Transconductance "Current Amplifier"

Thank you very much for posting your schematics here.

Encouraged by the substantial decrease in both harmonic and intermodular distortion I measured when driving several different drivers with a series resistor (20 to 90 ohms with a 300W class D amplifier) I tried to follow your instructions and wiring scheme. My idea is to drive my actively crossed upper midrange drivers with a current source amplifier way above the resonance frequency domain.
When connected to a 4 ohm dummy load with no input I measured 22V across the load and had to turn it off quickly due to the power dissipation - I used the light bulb trick (2 X 40W) - that lit up the room. :) Of course I could have made a mistake, but after several checks the wiring seems to be OK - not that any mistakes can ruled out by that of course.
You write one should null the output - is that the problem? And if so, could you - or someone else reading - this please help the chemist out? I'm very much a novice when it comes to circuitry.

Regards

I believe you null the output with the 10k potentiometer. Adjust it for minimum DC output.

Do you have the RC 0.1 uF + 22R across the speaker as shown in the schematic? These circuits can easily oscillate to full power at high frequencies.
 
I believe you null the output with the 10k potentiometer. Adjust it for minimum DC output.

Do you have the RC 0.1 uF + 22R across the speaker as shown in the schematic? These circuits can easily oscillate to full power at high frequencies.

Exactly.

Also, put a dummy load on the output first - a current output amplifier is unusual in that, if you have a DC offset at 4 Ohm, that offset will double at 8 Ohm and then double again at 16 Ohm and so on.

I use a dummy load resistor, adjust to near zero, then increase the resistor and fine tune further. With no load, it will be very touchy.

It sounds like the wiring is OK - it's just that current amplifiers behave very different to the norm.

Cheers, Joe
 
Last edited:
Exactly.

Also, put a dummy load on the output first - a current output amplifier is unusual in that, if you have a DC offset at 4 Ohm, that offset will double at 8 Ohm and then double again at 16 Ohm and so on.

I use a dummy load resistor, adjust to near zero, then increase the resistor and fine tune further. With no load, it will be very touchy.

It sounds like the wiring is OK - it's just that current amplifiers behave very different to the norm.

Cheers, Joe

Thanks for your circuit. I am going to give it a try as soon as some boards show up. Looking forward to experimenting. I will be using the LM3886. If I can get it to work I will post a pic and some measurements.
 
I believe you null the output with the 10k potentiometer. Adjust it for minimum DC output.

Do you have the RC 0.1 uF + 22R across the speaker as shown in the schematic? These circuits can easily oscillate to full power at high frequencies.



Exactly.

Also, put a dummy load on the output first - a current output amplifier is unusual in that, if you have a DC offset at 4 Ohm, that offset will double at 8 Ohm and then double again at 16 Ohm and so on.

I use a dummy load resistor, adjust to near zero, then increase the resistor and fine tune further. With no load, it will be very touchy.

It sounds like the wiring is OK - it's just that current amplifiers behave very different to the norm.

Cheers, Joe


The RC-filter is in place. Adding an extra filter across the dummy load (that has always been there - as described in my first post) didn't make any difference. Changing the IC didn't make any difference. I know what DC offset is, and I tried the minimum and maximum settings for the trimpot - with no significant change - before posting here - didn't think I had to mention that.

Since the term "output nulling" not "DC offset" is mentioned with a resistor value three magnitudes lower than the trimpot circuit value, I figured this couldn't have anything to do with the offset adjustment trimpot even though the words pointed to something similar.
I probably made a mistake, which I can't seem to locate - if the depicted schematic works. Does anybody know of a PCB for something similar?
 
Last edited:
The RC-filter is in place. Adding an extra filter across the dummy load (that has always been there - as described in my first post) didn't make any difference. Changing the IC didn't make any difference. I know what DC offset is, and I tried the minimum and maximum settings for the trimpot - with no significant change - before posting here - didn't think I had to mention that.

Since the term "output nulling" not "DC offset" is mentioned with a resistor value three magnitudes lower than the trimpot circuit value, I figured this couldn't have anything to do with the offset adjustment trimpot even though the words pointed to something similar.
I probably made a mistake, which I can't seem to locate - if the depicted schematic works. Does anybody know of a PCB for something similar?

I could not find a suitable PCB so I hacked a garden variety gainclone PCB to make it more of a breadboard. I ordered some to me fabricated. They should be here this week. Not sure if I can get to building it immediately.

What load do you have connected to the output?
 
No, there is no plans for a PCB as this is an amplifier that few will be making.

The schematic posted is also the correct layout if you follow it as a wiring diagram as well. Keep the transformer well away from the input as that is obvious and the rest should be straight forward.

Cheers, Joe

Joe, just found this current domain amp schematic and is there a parts estimate build cost for this excluding the case ?
Would it also work with planar magnetic headphones 50/60 Ohm ?
 
Joe, just found this current domain amp schematic and is there a parts estimate build cost for this excluding the case ?
Would it also work with planar magnetic headphones 50/60 Ohm ?

Haven't done a parts estimate build, but this is not an expensive build. As for compatibility with headphones, that I cannot say for sure, what impedance are they and have the published an impedance plot? Planars tend to have flat Z, a good thing. I would guess that a 15 Ohm 10W resistor on the output of the amplifier, would make many headphones compatible, and even tweak that resistor by ear and get away with the highest value possible. At least, it is less likely that this amp does any damage to the headphones in a way that many speakers could. So it is rather safer and I suspect this would make a rather above the norm quality wise headphone amp.

Cheers, Joe
 
Haven't done a parts estimate build, but this is not an expensive build. As for compatibility with headphones, that I cannot say for sure, what impedance are they and have the published an impedance plot? Planars tend to have flat Z, a good thing. I would guess that a 15 Ohm 10W resistor on the output of the amplifier, would make many headphones compatible, and even tweak that resistor by ear and get away with the highest value possible. At least, it is less likely that this amp does any damage to the headphones in a way that many speakers could. So it is rather safer and I suspect this would make a rather above the norm quality wise headphone amp.

Cheers, Joe


FostexT50RP 50ohm/LCD-2 60 ohm flat impedence.

LCD-2 maybe best with 3watts of headroom per ear .....so 4-6 watts max total output.

Well it is good to know that it may be possible as current domain headphone amps are few and far between, thus prices of $1800-$3000 which is priced on lack of competition more than exotic internals I'd guess.
 
music soothes the savage beast
Joined 2004
Paid Member
Hi Joe, any opinion about this circuit of current amp? I got plenty of LM1875, if it looks promissing, I may go ahead an built it, just for fun.
 

Attachments

  • trans amp LM1875.jpg
    trans amp LM1875.jpg
    27.3 KB · Views: 1,838
  • Like
Reactions: 1 user
Hi Joe, any opinion about this circuit of current amp? I got plenty of LM1875, if it looks promising, I may go ahead an built it, just for fun.

Why not have some fun then? That circuit, wonder what the DC levels would be like if copied just as shown. But a trimpot to the negative rail may be needed. That 220R on the output will hep when no load or go really high, but with load, needs to be trimmed.

Cheers, Joe
 
Just want to report...I built the circuit above, and not only its works fine just like it is, it sounds great. While normal recommended LM1875 circuit is ok, inverting with buffer sounds better. However, this transconductance circuit beats them all. It sounds liquid, like good tube amp.

Well done.

Yup! It surprised even a tube guy like me.

Cheers, Joe
 
Really interesting amps!

adason, are my calculation of gain of your amp right 22k/0.25+1 = 89 ???

O no, I just notice, R7 is 0.25 ohm, not K. How large is gain and how to calculate it?

Where you place R6 - close to chip or on speaker terminals?

You can see a discussion of current source op amp configurations here:
Variable Amplifier Impedance

The gain depends on the load impedance.

No current flows into/out of the + terminal so the 22k resistor will not affect the gain.