It's a pity this design has not been followed up in this forum. I suspect TNT's odd post may have chased it away.
However this idea has now developed into the headphone amp titled DCA02 - details can be found at KA-Electronics.com.
About 10 days ago I finalised putting one together and boxing it up.
I cannot recommend it too highly. It is now difficult for me to stop listening through it with my Focal Clear phones.
It is in direct competition with a Whammy that I knocked when it first surfaced and Richard Murdy's Sapphire amp.
I never liked the Whammy finding it plummy and airless so that was put to one side.
The Sapphire suits me better but by comparison the DCA02 is clean fulsome and sweet sounding with an amazing detailed delivery. Complex orchestral music is resolved in its entirety.
I have absolutely no connection with Wayne Kirkwood the designer by the way, but thought this thread is due for resurrection.
However this idea has now developed into the headphone amp titled DCA02 - details can be found at KA-Electronics.com.
About 10 days ago I finalised putting one together and boxing it up.
I cannot recommend it too highly. It is now difficult for me to stop listening through it with my Focal Clear phones.
It is in direct competition with a Whammy that I knocked when it first surfaced and Richard Murdy's Sapphire amp.
I never liked the Whammy finding it plummy and airless so that was put to one side.
The Sapphire suits me better but by comparison the DCA02 is clean fulsome and sweet sounding with an amazing detailed delivery. Complex orchestral music is resolved in its entirety.
I have absolutely no connection with Wayne Kirkwood the designer by the way, but thought this thread is due for resurrection.
I revived my Kirkwood THAT1646 HPA and connected it to my linear PSU with active bridges and LDO regulators. Then used an OPA1656 buffer with 14dB gain to drive the HPA. Sounds terrific. With +/-18v rails I wonder how many Watts this amp can deliver into 32ohm headphones? My estimate is 4.7W. If I made a balanced output version with double the amps, possibly 19w!? That’s sort of crazy of course.
Last edited:
A Simple 10W Direct-Coupled Class-A Power Amplifier
Thanks for the kind words Mildrew and xrk971.
Some years ago I played around with a 10W (8Ω) version. I used a pair of 18V laptop switchers to power it.
Dual Class-A 10W Heatsink Connections
Schematic of a Simple 10W Direct-Coupled Class-A Power Amplifier Using a DRV134, 2SA1962 and 2SC5242.
A Simple 10W Direct-Coupled Class-A Power Amplifier - Pro Audio Design Forum
The schematics for the DCAO-II can be found here: Dual Class-A II Line and Headphone Output Construction Information - Pro Audio Design Forum
I don't offer the 1646/DRV134 boards any longer but I think the DCAO-II could be used to extend the power output in full class-A. I've never tried it though.
For driving speaker loads I do recommend the DRV134 due to its cleaner clipping characteristic when the common mode drive of this circuit is used. Note the addition of the R1/R2 voltage divider required for the DRV134. The 1646 has a differential (bias voltage) gain of "2" the DRV134 has a differential gain of "4."
An interesting side story: Once these images hit the web I got contacted by a prop master wanting to buy the Augat heatsinks. Turns out these heatsinks were used in the original Ghostbusters back packs.
Thanks for the kind words Mildrew and xrk971.
Some years ago I played around with a 10W (8Ω) version. I used a pair of 18V laptop switchers to power it.
Dual Class-A 10W Heatsink Connections
Schematic of a Simple 10W Direct-Coupled Class-A Power Amplifier Using a DRV134, 2SA1962 and 2SC5242.
A Simple 10W Direct-Coupled Class-A Power Amplifier - Pro Audio Design Forum
The schematics for the DCAO-II can be found here: Dual Class-A II Line and Headphone Output Construction Information - Pro Audio Design Forum
I don't offer the 1646/DRV134 boards any longer but I think the DCAO-II could be used to extend the power output in full class-A. I've never tried it though.
For driving speaker loads I do recommend the DRV134 due to its cleaner clipping characteristic when the common mode drive of this circuit is used. Note the addition of the R1/R2 voltage divider required for the DRV134. The 1646 has a differential (bias voltage) gain of "2" the DRV134 has a differential gain of "4."
An interesting side story: Once these images hit the web I got contacted by a prop master wanting to buy the Augat heatsinks. Turns out these heatsinks were used in the original Ghostbusters back packs.
Last edited:
Thank you.
It was 0.0449% THD if I'm looking at the right post. The THD+n numbers on those screen caps don't look right to me.
Later in the thread, after adding the switchers for ±18V, I did an 8W 8Ω FFT showing 0.019% THD with 1200 mA Iq.
I need to revisit this project with more modern instrumentation.
A Simple 10W Direct-Coupled Class-A Power Amplifier - Page 8 - Pro Audio Design Forum
Not shown in the photos are the thermal sense transistors. Those are on the underside heatsink and are shown in the original thread.
Dual 10W Direct-Coupled Class-A Power Amplifier Powered by Meanwell 60W "Laptop" Switching Power Supplies.
I've never pursued it, but my hunch is that one of the factors that contributes to the output stage linearity - in addition to high class-A current - is the stiffness of the Vbe spreader.
The Vbe spread is maintained by DC feedback in addition to thermal.
In the case of the THAT1646/DRV134 the feedback is internal to the device.
It was 0.0449% THD if I'm looking at the right post. The THD+n numbers on those screen caps don't look right to me.
Later in the thread, after adding the switchers for ±18V, I did an 8W 8Ω FFT showing 0.019% THD with 1200 mA Iq.
I need to revisit this project with more modern instrumentation.
A Simple 10W Direct-Coupled Class-A Power Amplifier - Page 8 - Pro Audio Design Forum
Not shown in the photos are the thermal sense transistors. Those are on the underside heatsink and are shown in the original thread.
Dual 10W Direct-Coupled Class-A Power Amplifier Powered by Meanwell 60W "Laptop" Switching Power Supplies.
I've never pursued it, but my hunch is that one of the factors that contributes to the output stage linearity - in addition to high class-A current - is the stiffness of the Vbe spreader.
The Vbe spread is maintained by DC feedback in addition to thermal.
In the case of the THAT1646/DRV134 the feedback is internal to the device.
Last edited:
Hi Mediatechnology,
I think TTC5200 and TTA1943 output devices would work fine here then? If the DRV134 has an on board zener diode supply, 19v laptop bricks can be used then? Would the addition of bootstrap caps permit operation closer to rail to rail? I think output DC protection SSR might be a good idea for a DC coupled amp like this.
Thanks,
X
I think TTC5200 and TTA1943 output devices would work fine here then? If the DRV134 has an on board zener diode supply, 19v laptop bricks can be used then? Would the addition of bootstrap caps permit operation closer to rail to rail? I think output DC protection SSR might be a good idea for a DC coupled amp like this.
Thanks,
X
Hi,
I hurriedly threw some components on this PCB (project). Don't laugh at improvisations.
I connected it to my laptop(Spotify) and listened to it with the AKG702. The supply voltage is 12.5V and I am going to increase it.
Yes, I need a little more gain.
I guess I don't have too many possibilities with this scheme and I would make a preamplifier with 5532 ... I think it's not a very good idea and maybe I'll get some advice.
I hurriedly threw some components on this PCB (project). Don't laugh at improvisations.
I connected it to my laptop(Spotify) and listened to it with the AKG702. The supply voltage is 12.5V and I am going to increase it.
Yes, I need a little more gain.
I guess I don't have too many possibilities with this scheme and I would make a preamplifier with 5532 ... I think it's not a very good idea and maybe I'll get some advice.
Attachments
It’s a great sounding amp. Very neutral and clear. Not fatiguing at all. Powerful as heck.
You can get the board from KA Electronics which is Wayne Kirkwood’s store, or you can also use the layout made by Prasi and order your own boards. For about the same price as ordering your own, I would support Wayne’s store.
Dual Class-A II Reference Headphone Amp Bare PC Board
You can get the board from KA Electronics which is Wayne Kirkwood’s store, or you can also use the layout made by Prasi and order your own boards. For about the same price as ordering your own, I would support Wayne’s store.
Dual Class-A II Reference Headphone Amp Bare PC Board
Last edited:
Thanks for the kind words everyone.
For those wanting to build the original version do heed post #9 https://www.diyaudio.com/forums/hea...headphone-amp-using-that1646.html#post2843960
The THAT1646 and TI DRV134 can both be used but the DRV134 requires scaling the Vbe reference by 1/2. If a DRV134 is plugged into a 1646 socket the bias current will be far too high.
For those wanting to build the original version do heed post #9 https://www.diyaudio.com/forums/hea...headphone-amp-using-that1646.html#post2843960
The THAT1646 and TI DRV134 can both be used but the DRV134 requires scaling the Vbe reference by 1/2. If a DRV134 is plugged into a 1646 socket the bias current will be far too high.
Thanks! I enjoy building Vectorboards and still do it occasionally for simple one-of-kind-things.
IIRC that FFT must have been a null test due to the -70 dB fundamental. That was a very long time ago.
I now have "ULDO-Nacho" to generate ultra-low THD tone and notch the DUT output. ULDO's THD is <<-140 dBc, the lowest I've measured from it is around -151 dBc. With ULDO-Nacho I don't have to null test. ULDO-Nacho-Atto Ultra Low Distortion Oscillator, Notch Filter and Attenuator Construction - Pro Audio Design Forum
Using ULDO-Nacho the THD of the dual class-A II delivering 500 mW into 30Ω is coming in at around -112 dBc, 2.3 ppm or 0.00023%. This was with overall NFB on and operating in class-A/B. Dual Class-A II Line and Headphone Output Construction Information - Pro Audio Design Forum
IIRC that FFT must have been a null test due to the -70 dB fundamental. That was a very long time ago.
I now have "ULDO-Nacho" to generate ultra-low THD tone and notch the DUT output. ULDO's THD is <<-140 dBc, the lowest I've measured from it is around -151 dBc. With ULDO-Nacho I don't have to null test. ULDO-Nacho-Atto Ultra Low Distortion Oscillator, Notch Filter and Attenuator Construction - Pro Audio Design Forum
Using ULDO-Nacho the THD of the dual class-A II delivering 500 mW into 30Ω is coming in at around -112 dBc, 2.3 ppm or 0.00023%. This was with overall NFB on and operating in class-A/B. Dual Class-A II Line and Headphone Output Construction Information - Pro Audio Design Forum
Hi Wayne,
That’s great stuff there on your ULDO-Nacho! You have been busy. I like how your design is made for balanced output. Is it available preassembled?
For simple circuits I will sometimes use a black Sharpie as acid resist and etch a board quickly:
This was the DLH Class A amp by DiegoMJ:
Last edited:
Hi,
first of all i would like to thank you all Wayne, Xrk, Prasi..
still trying to source ICs but the other day the pcbs arrived Prasi designed;
i have a question abot signal input labels on X1 terminal where the IN labeled trace goes to the GND on X3 terminal..
shouldn't it be in reverse labeling (IN/G) or am i missing something?
thanks in advance..
Attachments
Hi Wayne!
I've started to cook up a small 60x50 board inspired by your design! I have tried to keep it as simple and sparse as possible. I hope to run a 16ohm compression driver with this little amp.
Cross coupled THAT128x
Dual THAT1686 in 'BTL' mode
DMMT5551/DMMT5401
OP27 servo (maybe)
Very much WIP. I have made a start simulating in Tina-Ti. I'd be happy to hear your comments.
Let me start a new thread so not to clutter this one up with this version.
I've started to cook up a small 60x50 board inspired by your design! I have tried to keep it as simple and sparse as possible. I hope to run a 16ohm compression driver with this little amp.
Cross coupled THAT128x
Dual THAT1686 in 'BTL' mode
DMMT5551/DMMT5401
OP27 servo (maybe)
Very much WIP. I have made a start simulating in Tina-Ti. I'd be happy to hear your comments.
Let me start a new thread so not to clutter this one up with this version.
Last edited:
Nice!
Where is the third transistor/Vbe multiplier? It provides thermal feedback. I see the "bias" port on the sch but no circuit.
You don't really need a servo in this circuit. The output Vos is low - it's the sum of the 1646 output offsets plus 1/2 the Vbe difference of the NPN/PNP output pairs. Typically its under 5 mV.
I would remove C2 and C8.
Where is the third transistor/Vbe multiplier? It provides thermal feedback. I see the "bias" port on the sch but no circuit.
You don't really need a servo in this circuit. The output Vos is low - it's the sum of the 1646 output offsets plus 1/2 the Vbe difference of the NPN/PNP output pairs. Typically its under 5 mV.
I would remove C2 and C8.