Dual Class-A Line and Headphone Output Board Construction Information

Construction information for DIY projects, including the MS Mid Side Matrix, Elliptic Equalizer, Mastering Console, Phono Transfer System, Insert Switcher and the Dual Class-A Amplifier. You can post your baby pictures here.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Dual Class-A Line and Headphone Output Board Documents

Post by mediatechnology »

Paul -

I've had a customer with this issue and have seen this myself.
In both cases the pot was mis-wired because the silkscreen created confusion at the Phoenix connector.
"I" and "O" refer to the circuit inputs and outputs, not the pot's In and Out.

Check that the "O" Phoenix terminal is to the end terminal of the pot. (Terminal 1 on the TKD.)
The "I" Phoenix terminal is to the wiper. (Terminal 2.)
I think ground on the TKD is marked "C."

With a 50K pot you might want to add a coupling cap due to bias current.
The value is a little high for the load impedance of the 1646 and may bow the taper but that's not your immediate problem.
Gold
Posts: 676
Joined: Sat Dec 01, 2007 5:20 pm

Re: Dual Class-A Line and Headphone Output Board Documents

Post by Gold »

mediatechnology wrote: "I" and "O" refer to the circuit inputs and outputs, not the pot's In and Out.
Ahh, that sounds right. I misunderstood the labelling. I should have double checked the schematic. I probably did just that for the other build and figured "I got this" for this one. Nope. Thanks.
Gold
Posts: 676
Joined: Sat Dec 01, 2007 5:20 pm

Re: Dual Class-A Line and Headphone Output Board Documents

Post by Gold »

Yep, that was it.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Dual Class-A Line and Headphone Output Board Documents

Post by mediatechnology »

Been there myself: viewtopic.php?f=6&t=745
If you want to skip to the punchline its the last post. :oops:
Gold
Posts: 676
Joined: Sat Dec 01, 2007 5:20 pm

Re: Dual Class-A Line and Headphone Output Board Documents

Post by Gold »

mediatechnology wrote:
With a 50K pot you might want to add a coupling cap due to bias current.
The value is a little high for the load impedance of the 1646 and may bow the taper but that's not your immediate problem.
I must admit I don't really know how to look at the DRV134 for input impedance because the usage is non standard. With a 10k pot the maximum source impedance would be 2k5Ω. Even that seems high as a source impedance for a 1646 (DRV134). A 50k pot would present a maximum source impedance if 12k5Ω.

How would I calculate the value of a coupling cap?
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Dual Class-A Line and Headphone Output Board Documents

Post by mediatechnology »

I have to look at the input impedance of a DRV134 in that configuration.
When I looked at that it was for a 1646 at around 4K minimum.
I think the DRV134 is actually higher.

My guess is 47 uF would be a good value based on the 1646.
But if the bias current DC isn't audible I wouldn't add the coupling cap.

FWIW: A good RC calculator can be found here: http://www.muzique.com/schem/filter.htm

The 1646/134 do require a low source impedance when they're used in the "conventional" balanced output configuration.
In this topology the source impedance doesn't degrade performance like it would in a balanced output.
bobkatz
Posts: 11
Joined: Tue Jul 11, 2023 10:07 am

Re: Dual Class-A Line and Headphone Output Board Documents

Post by bobkatz »

I know the DCAO_173 board has long been superceded! But I finally got around to building a headphone amp with one of these and I thought I would report on the performance and sound. Since it keeps on improving, what can you lose getting the next generation!

Also important to note that I'm using a very high pedigree power supply, the SIGMA 22 from AMB audio, which my assistant built under my direction. The Sigma is an all-discrete tracking supply which could throw some sparks if you want, but you don't since it's not well protected from transient shorts! I think half of the sound of any amplifier is due to the power supply quality.

I'm regulating to +-15 VDC, by the way, as recommended by Sir Wayne. I biased it WAY above the 85 ma (280 mv. at Test points) described in this thread by Wayne. I pushed it to a scary 409 mv! That's 123 ma. I chose this bias based on watching the third harmonic of 1 kHz go down till it was pretty much as far down as it would go. I think that reducing the third harmonic will increase the perceived transparency of the amplifier.

Using my Fluke temperature probe on the heat sinks and/or on the nuts holding the output transistors in place, the highest temperature I could measure was an excellent 56 degrees C! I guess I like to take risks, as I'm using a closed case with no ventilation. After 15 minutes with the case cover closed, I got the 56 degrees. So I think we are gonna be pretty safe with this bias.

Next, the measurements, attached.
Attachments
Kirk Freq response.pdf
(323.06 KiB) Downloaded 357 times
Kirk IM.pdf
(243.02 KiB) Downloaded 376 times
Kirk 1k +9 dBu 20R load.pdf
(242.84 KiB) Downloaded 361 times
Kirk 1k 0 dBu 20R load.pdf
(244.29 KiB) Downloaded 403 times
Last edited by bobkatz on Tue Dec 05, 2023 6:13 pm, edited 1 time in total.
bobkatz
Posts: 11
Joined: Tue Jul 11, 2023 10:07 am

Re: Dual Class-A Line and Headphone Output Board Documents

Post by bobkatz »

Measurements description:

The Prism Lyra ADC/DAC I use for measurement has a noise shaped noise floor with an HF rise from 50 through 96 kHz, the limit of my measurement, which I take at 192 kHz sampling and Spectraplus software.

Frequency response: I use a stepped 1/3 octave tone, which explains the jaggies where you might expect a smooth curve. It's 0.0 dB as low as 20 Hz, -1 dB at 10 Hz!!! And rolls off rapidly below 10 Hz. Flat 0.0 dB to 20 kHz. Minus maybe -0.2 dB at 30K. Minus 2 dB at 60 K and rolls off rapidly above that.

1 kHz harmonic distortion with a 20 ohm load with 1 kHz at a little higher than -1 dBu. This would produce about 106 dB SPL in the Audeze LCD-5 phones, whose impedance is listed as 14 ohms. Second harmonic is at about -90 dBu and third about -97 dBu. That's all I can see, quite pure with no other traces of harmonic distortion.

Pushing the amp 9 dB harder, still with a 20 ohm load, to around +9 dBu at 1 kHz. Second harmonic raises to around -72 dBu, third to about -68, and now we can see 4th harmonic at around -100 and 5th at around -92. Both channels are very symmetrical but I see traces of higher order harmonics in the right channel (below -120 dBu) that are clean in the left. This would be VERY loud in any modern low impedance/sensitive headphone. I did not test the amp with a high impedance phone like the Sennheisers so I cannot comment on its performance in that case.

IM distortion, with 19 kHz and 20 kHz at a 1:1 ratio, both of those outputting at about -9 dBu. Pretty good performance. I've seen much worse and VERY rarely better. 1 kHz difference product at -100. HIgh frequency side bands apparently at 18kHz and 21 kHz also at -100, and harmonics of these also around -100 around 40 kHz.

Voltage gain is about 18 dB with a THAT 1240 input. Perhaps the 1246 would be more appropriate as 18 dB is a lot of gain. But it's working fine with about a 20 dB dithered digital attenuation and my Prism Callia DAC, which produces -14 dBu at 0 dBFS. If you have a nominal "+4 output" device, then for sure you would need the 1246 input.

Crosstalk at 1 kHz with a 20 ohm load is not fantastic. I measured about 59 dB. Something about the current flow through a Left/Right channel headphone jack that has a common ground. It's close to 90 dB with no load, so the crosstalk is not due to any design error.
Last edited by bobkatz on Tue Dec 05, 2023 6:19 pm, edited 2 times in total.
bobkatz
Posts: 11
Joined: Tue Jul 11, 2023 10:07 am

Re: Dual Class-A Line and Headphone Output Board Documents

Post by bobkatz »

Finally, the sound. I listened to a number of my own reference recordings using the Audeze Flagship LCD-5 phones, which are revealing yet quite pure, and using Bob's LCD-5 EQ that has been officially endorsed by Audeze. The EQ is implemented in the digital domain.

I matched the Kirkman amp subjectively in level to the superb Audeze Deckard amplifier, which is listed as "Sold Out", but it's really discontinued. It's discontinued I think because they couldn't make any money at the $700 price they decided to let it go for. The Deckard is a remarkable all class A Audiophile discrete amplifier, a worthy contender for sure. It's claimed to be able to produce up to 4 Watts! Which I confirmed long ago. I like its sound a lot -- despite all the contenders I've put up against it, up to BEFORE the Kirkman, the Deckard has bested (is that a word?) them all.

Let's see, the Kirkman never reached 1% THD, it was more like in the 0.06% range if I recall, at these levels! So I have no idea of the official clipping point. Anyway, I think 1% is a terrible standard for clipping or max peak rating because most solid state gear sounds totally ratty and unusable at that point. But if we call the +9 dBu point into 20 ohms as its recommended max, that would be 242 mW, (which isn't very much power in the rarified audiophile headphone world). But let's be realistic: The LCD-5 is spec'ed at producing 90 dB SPL with 1 mW. So with 242 mW would produce 113 dB SPL on peaks. And it's still at that point well below 1% THD.

After matching the Deckard against the Kirkman, there were extremely subtle differences. Both amps sound robust, full, warm, clear, deep and transparent! Both amps were able to play satisfyingly loud. I am quite pleased with the Kirkman and the differences are a matter of nuance. PERHAPS the Deckard sounds more transparent, but I also think the Kirkman sounds just a little clearer (clearer in the sense of accurate, not artificially bright) at the top end. I'm not 100% sure as the differences are so subtle I know I could not pass a blind test on these. That's how good the Kirkman amp sounds to me.
User avatar
mediatechnology
Posts: 5437
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Dual Class-A Line and Headphone Output Board Documents

Post by mediatechnology »

Bob, thank you for posting this!

I just looked back and you bought the assembled board from me on September 15, 2015.
USPS Priority Mail was only about $6 then.

You and the late Joe Neil were the first two customers in my store.
My how time flies...

I'll take a look at the data and figure out how may mW you can get at the Iq setting you have and remain in class-A.

One day just for fun I drove a pair of Minimus 7 with this 100 mW headphone amp and it was way out of class-A.
I was amazed at how good it sounded considering that at the level I was listening to, maybe a Watt, it was measuring around 1% THD.
viewtopic.php?t=679
Post Reply