OTL Headphone Amp project.

[DrNomis_44]

Update:

Just got the last lot of caps and resistors I ordered from Mouser Electronics, delivered to my front door a few minutes ago, just have to wait till they send the 43k/2W resistors I ordered since they were on back-order, all that's left to do now is to order the new Hammond 369AX power transformer that I need from Evatco and then I'll be able to start work on building the new 6N1-P OTL Headphone amp.

[Simon Barden]

As I said, the 43k is a typo and the design is for a 42k resistor. It probably won’t have a huge effect on the bias, just your wallet.

[DrNomis_44]

As I said, the 43k is a typo and the design is for a 42k resistor. It probably won’t have a huge effect on the bias, just your wallet.

I had actually ordered 43k/5%/2W resistors, so since they are 5% types they will measure anywhere within +/- 5% of 43k, on the low side they could actually measure:

43,000-5%= 40,850 or 40.85k

And on the high side:

43,000+5%= 45,150 or 45.15k

But note that in each case, that's for a worst-case scenario.

Hopefully my maths is correct, but I'm happy to be proven wrong.

[DrNomis_44]

I will also need to order two dual 50uF/500V can caps from Soundlabs Group, along with the new power transformer.

[DrNomis_44]

Just out of curiosity, I did some calculations to see what values of resistances I would get if I put two garden-variety 82k/5%/1W carbon film resistors in parallel, here's what I would get in the worst-case scenarios.....


82,000-5%= 77,900 or 77.9k

Two 77.9k resistors in parallel would give us:

77,900/2= 38,950 or 38.95k

On the low side.

82,000+5%= 86,100 or 86.1k

two 86.1k resistors in parallel would give us:

86,100/2= 43,050 or 43.05k

On the high side.

[Simon Barden]

You could afford to get a bunch if the common 82k resistors and test them with a meter and pick the best ones to give you close to 42k. Or 41k or 43k. Whichever value gives the best results for biasing.

[Simon Barden]

Another option would be a 39k 2W resistor in series with a 2W wirewound 10k multi-turn resistor (about $7 each here). You can then easily trim the bias resistance for optimal results and you don’t need to worry about tolerances.

[DrNomis_44]

Another option would be a 39k 2W resistor in series with a 2W wirewound 10k multi-turn resistor (about $7 each here). You can then easily trim the bias resistance for optimal results and you don’t need to worry about tolerances.

Yeah, that would be a good alternative, I'd need to do some Ohms Law calculations to see exactly how much power the 10k multi-turn adjustable resistor would be dissipating, I wonder how much a standard blue-coloured vertical-mount ten-turn 10k trimpot can dissipate before it burns-up? real-estate in the chassis is a bit limited, I've had to play-around with the layout a bit to see if I could get everything to fit in it, I'm using some 8-way tagstrips to mount most of the components, and I'll be using some stranded wire to wire from the tagstrips to the valve sockets.

I might do some Ohms Law calculations to find out exactly how much power the anode resistor of the first triode gain stage is dissipating, I'd have to work out the voltage-drop across it, and the maximum current flow through it though, if I take the anode voltage of the first triode gain stage, 172V, away from the + HT supply voltage, 348V, we get the following:

348-172= 176V

So, that 41k/2W anode resistor should have a voltage drop of 176V DC across it by my calculations.

So, if we apply Ohms Law to work out the current flow in that 41k/2W resistor we get the following:

I=E/R

I= 176V/41,000

= 0.0042926829268293

0.0042926829268293 X 1000= 4.292682926829268

Or about 4.29 milliamps if my calculations are correct.

If we do the power dissipation calculation according to Ohm's Law, we get this:

P (in watts)=I X V

P= 4.292682926829268 X 176V

= 755.5121951219512

755.5121951219512/1000

= 0.7555121951219512 watts.

So therefore that 41k/2W anode resistor isn't even dissipating the full power it is capable of dissipating.

[DrNomis_44]

Another option would be a 39k 2W resistor in series with a 2W wirewound 10k multi-turn resistor (about $7 each here). You can then easily trim the bias resistance for optimal results and you don’t need to worry about tolerances.

Since the needed 42k/2W resistors are going to be hard to order online in small quantities, I'm going to buy four 82k/1W carbon-film resistors from my local Jaycar Electronics store next week, and give them a try in my 6N1P OTL headphone amp build and see what results I get, I can always desolder them and swap them out with two 42k/2W metal film resistors at a later date, I'm going to compile a list of the parts I do have so I can work out what parts I need to order online this year, once I've got the 6N1P OTL amp build up and running and working properly, I'm going to do some tests on it using my diy XR2206 Function Generator, my DSE Q1804 20MHz Dual Trace Scope, and two 33 Ohm/1W carbon film resistors which will be used to, roughly, simulate a pair of 32 Ohm headphones plugged into the amp, so I can view the output waveforms the amp produces when driving 32 Ohm headphones, I'll post some screenshots while I'm at it.

While I'm doing the tests on my OTL headphone amp build, I'll see if I can set the output level of the diy XR2206 Function Generator to the output level of the Digitech D/A converter I'll be using to convert the optical-out of my 43 inch Blaupunkt TV to stereo analog line-outs, mainly so I can evaluate how the amp is going to perform, if necessary, I'll be able to make adjustments to the resistor values to maximise the performance, that's going to be the most fun part of the build-project, seeing if I can get the best performance out of it, its well-known that valves of a particular type tend to vary quite a bit due to manufacturing tolerances, so because of that there's no way to tell how well my three Svetlana 6N1P valves will perform in practice compared to the 6N1P valves used in the original design, plus the designer used some circuit simulator software in their design process, and the valve simulations in the software a usually idealized for consistency, it will be very interesting to see how well my build performs compared to the original design.

[JohnH]

Good to see you're still at it doc!