RJM Audio Sapphire Desktop Headphone Amplifier

[rjm]

@lumlum

The build looks very nice. The issue is in both channels so it is a systematic error.

The first thing to check after confirming the power supply voltages is the output bias currents.

You'll need to measure the voltage drop across R17 and R18, and compute the current as I = V/R.

It's possible you mixed up the transistors, so the circuit is not providing amplification. Or it's possible you mixed up resistor values, so the circuit gain is less than expected.

I note you are running in open loop mode, using R3. What is the value of R3 and R2, and what headphones are you using?

 

[lumlum]

I'm starting to need glasses.🤓
They actually gave me 4.75k instead of 4.75R for R17-R20.

Thanks Richard, I don't know how many times I would have looked over there without seeing it.

 

[rjm]

At least you used Dales, so you can see the value at a glance.

Replacing those should get you back on your feet.

 

[lumlum]

This makes the mistake all the more stupid, because I've gotten used to aligning the value upwards.
And I usually check the resistors with the DMM before soldering them in, which I didn't do this time either.
Unfortunately, I don't get a Dale with a value of 4.75ohms quickly ordered, except from Mouser and there it doesn't pay off because of the high shipping costs.
I have now ordered a simple Vishay MBB0207 4.7ohm.

Yes Richard, then my Sapphire will work.
I'm very excited because i could hear that the Sapphire has a lot of potential.

 

[EUVL]

Beyschlag MMB0207 is just as good.
But if you must have Dale, then :
Vishay Dale CMF55 4,7 Ohm, 5% :: 0 Ohm - 99 Ohm :: CMF55 :: Vishay Dale :: Resistors :: Passive Components :: Electronic Parts :: Banzai Music GmbH

(A bit pricey, but you only need a couple.)


Cheers,
Patrick

 

[lumlum]

Beyschlag MMB0207 is just as good.
But if you must have Dale, then :
Vishay Dale CMF55 4,7 Ohm, 5% :: 0 Ohm - 99 Ohm :: CMF55 :: Vishay Dale :: Resistors :: Passive Components :: Electronic Parts :: Banzai Music GmbH

(A bit pricey, but you only need a couple.)


Cheers,
Patrick

No, it doesn't have to be Dale.
So I chose the MBB series from Vishay, and they have already been ordered.

But many thanks anyway for the link Patrick.

 

[EUVL]

I also use Beyschlag MBB all the time.
Available also from Buerklin.


Patrick

 

[lumlum]

They are also good resistors and I also like to use them.
I ordered it from TME on Wednesday and received it yesterday.
Of course I also soldered it in yesterday, but I couldn't hear it anymore.
I'll catch up on that tonight, or rather tonight as I know myself.

 

[lumlum]

I've been listening to my sapphire for a while now, and I'm thrilled.
I already thought that this amplifier would be top, but I didn't expect such an amazing result.

Since I listen to electronic music, I like the resolution and the large space, and above all the powerful, deep bass.

Thank you Richard for your excellent work you are great.

A few more pictures of the circuit boards with the changed resistances for R17-20

 

[rjm]

Kudos for a job well done.

 

[rjm]

I've been fielding questions from people wanting to build balanced versions of the Sapphire.

A. Yes, it is possible. Just remember that the Sapphire circuit is just a gainblock to provide voltage amplification and current gain. You're are responsible for the rest of the signal path. For balanced output, that means a phase splitter. For balanced input, that means a summing circuit. Both of those functions can be achieved with line transformers (Jensen, Lundahl) or a couple of dual op amps.

Two alternatives not pictured:

1. Use an existing balanced line preamp to replace the front end depicted in the second image. Since the volume control is already taken care of, the line output from the preamp just connects directly to the four Sapphire boards.

2. If you don't propose to include balanced inputs, its possible to leave out the two conversion stages and run balanced from end-to-end (with a four deck volume control).

 

[jwags81818]

Running your input wiring right past the transformers without any shielding??

This is the new discussion thread for the Sapphire amplifier.

Previous threads,

183588-headphone-amplifier-drop-replacement-phonoclone-3-vsps300

160460-rjm-audio-gainbloc-preamplifier-mk-i

are redirected here.

Information about the kit group buy is at this thread.

 

[Stan63605]

Hello, Richard, I have been happily playing Sapphire 4.0 since 2017. Egad, time passes. 4.0 was a prototype if I remember correctly. I've just acquired upgraded cans - Beyerdynamic DT1990 Pro, 250 Ohms, 102dB SPL, and I'd like to adjust my set, but I'm unable to find the original BOM. 4.1 is apparently different enough I'm not sure what to change. I'd be grateful if you can dredge up the BOM for 4.0. You might recall that I was the first to assemble and test this variant and we had a surprise on one pcb, corrected with an addition, and it morphed into 4.0, which you did send - but my Sapphire seems to be unique. I attach images to refresh if that helps.

 

[Stan63605]

Reading through all the 4.1 stuff and your comment a couple months ago, I think I want to change R2 from 470R to 680R. Is that enough? I doubt that I'll ever be using 600 Ohm headphones, but I shan't be likely to use 32 ohm.

 

[rjm]

Running your input wiring right past the transformers without any shielding??

Probably refers to this photo, I guess. Yes I did. It was more reasonable in a previous build when the transformers were smaller. I upgraded the transformers to Triad VPM and things got really tight back there.

Not ideal. Neither is bringing the AC forward to the front panel switch and back again to the transformer primary.

However, you can get away with things in a headphone amp that you could not with a MC phono preamp. Turns out neither are a problem.

 

[rjm]

Hello, Richard, I have been happily playing Sapphire 4.0 since 2017. Egad, time passes. 4.0 was a prototype if I remember correctly. I've just acquired upgraded cans - Beyerdynamic DT1990 Pro, 250 Ohms, 102dB SPL, and I'd like to adjust my set, but I'm unable to find the original BOM. 4.1 is apparently different enough I'm not sure what to change. I'd be grateful if you can dredge up the BOM for 4.0. You might recall that I was the first to assemble and test this variant and we had a surprise on one pcb, corrected with an addition, and it morphed into 4.0, which you did send - but my Sapphire seems to be unique. I attach images to refresh if that helps.

Hi Stan, nice to hear from you. Yes, you were my guinea pig on the Sapphire 4 development, and a very nice guinea pig you were too!

4.0 was the same circuit as 4.1/4.2, but without some of the convenience features like the gain switch or the option to run closed loop. Looking over the schematics, the operating points did not change so nothing needs to be updated. You can change the gain by adjusting the combination of R2 and R3 in the same way as described in the current BOM's gain calculator (open loop), which suggests 17 dB gain (call it 15 dB ...) for the 250 ohm DT1990 Pro.

Keeping R3 as 10k, this means R2 should be about 1k. Which was, I believe, the value originally specified. If you need more gain, you can make R2 a little smaller.

I won't post the old BOM here to avoid confusion and because things in there might be out of date of flat out incorrect. I do have it, and all the other resources for the older versions archived on my PC, so please ask by email if you would like any old documents.

 

[Stan63605]

Hi, Richard. Yesterday I found BOM 4.2b and, compared the several versions in my file, and I thought perhaps I should switch R2 to 1K (rather than 680R), and then last night I read your #1716 note which tells me I was on the right track, thank you. I am intrigued by 4.2b with its giant electrolytic capacitors, whereas all the earlier versions had the more numerous tall, slim 1000uf/25V Nichicons. Is there any appreciable benefit to the bigger caps? Regardless, I decided that its time I must build the 4.1m pcbs with its switchable gain option. I'd actually purchased what I need to build it some time ago but 4.0 was working so well that the new build kept slipping down my job jar! Will fix that.

For five years I have been listening to DT880 Pro, 250R and I had accepted one reviewer's opinion that he could not reliably tell the sound between his DT880 and Beyerdynamic more expensive Tesla drivers. I think I need to find a better tester! The DT1990 Pro cans are quite wonderful and I immediately noticed their sound improvement. But the higher sensitivity of the DT1990 made clear that the click-steps in my volume control are now too coarse.

 

[Stan63605]

Interestingly, in my prototype 4.0, when I opened it up I saw that R3 is set at 12.7K, R2 is at 300R. I cannot find my notes, and I haven't looked through the 1700+ notes in the Sapphire forum in DIY! I think I'll reset R2 at about 600R and see how that works, and get to work on my 4.1b.

 

[Stan63605]

I fixed R2 at 820R, adjusted the DC offsets at 25mV. My volume control is at about 9:00 at reasonable levels. Perhaps I could have gone higher, 1K? Next time!

 

[rjm]

300/12.7k would have set the gain fairly high, probably closer to 25 dB than 15 dB. There isn't really a lot going on here, technically speaking, it's just a question of how loud you want the music to be at what volume level.