Let's Build a Phono Stage

I am still game , pm incoming.. dB

 

Boards ordered

 

PM sent !

 

Here's the Mouser BOM and gerbers. I won't be using this for my own build, as I have a bit of a stash to burn through, so please check it against the boards that you order.

What's included:

- All parts required to populate one board/channel
- Optional electrolytic caps
- Standard through-hole JFETS (J113). When four are used, this produces a gain of about 58dB at 20Hz, which should be OK for standard MM carts or in systems with large amounts of downstream gain.
- Heat sinks.

What's not included:

- Power supply. This CUI part is more than sufficient to power two channels.
- Optional surface mount JFETs. These develop about 3x as much gain, and would be a good option for less sensitive systems. 2SK3557 or CPH3910 (for a bit more gain, and even less noise) would be good choices.
- Cart loading. There are four spots on the board that can be populated by 1/4W resistors or mica capacitors. I recommend selecting resistors from Yageo's MFR series. A good option could be to use 4x 100k resistors, allowing 25k, 33k, 50k, 100k, and 1M by switching them in parallel, or to use a couple of Rs and a couple of Cs. Builder's choice.
- Heatsink mounting kits (6-32 bolts, insulators, thermal compound), wires, and hardware. The parts will not get terribly hot, so it's reasonable to pass on thermal compound.

Each channel is set to draw about 80mA. No part matching is required.

 

Thanks for all the work you have put into making it easy for us. My boards are finalized and should be shipped soon.

 

This has taken a little longer to get to than I had expected.

a brief progress report.

PSU sections built and tested.



- both PSUs power up as designed.

Boards assembled , cleaned and waiting to dry




edit - forgot about the JFET adapters = this will allow all 3 JFET options to be tested



still more to come ..dB

 

You're faster than me! The JFET adapters are a really clever solution, and it's great to see your build come together.

I noticed you left one of the jumpers unpopulated. This allows you to lift the ground between the power supply and circuit-- in hindsight, not a great feature. Unless you have a very specific use case, make sure you install it, or the circuit will not power on, and it will be frustrating to troubleshoot.

 

my plan was to add a small connector so that I could easily connect / disconnect the ground plane but if it is better to default to connecting it, I will just solder in a solid jumper.

I added the ability to change the CCS resistor as you recommended too

( and sharp eyes might notice C5 is unpopulated - I found the baggie with caps on the floor after I finished - so populated them )

If anyone goes down the path of the little adapters = I just bought these off amazon, much easier than making my own but the jfets had to be mounted backwards for the pinout to match the board layout - just an FYI to check the pinout vs adapter but I believe the most recent board files have the option for mounting the smd on the board

 

Quick update : both boards tested for power up and working. 63mA across the ccs - tested with J113 Jfets.

Now to find time for some case work, and the real test with music.

This is a well laid out board, easy to assemble. The only change I might make is to set the spacing for the power/input /output to accept various types of connectors. As designed it is set up for bare wire. The output might accept molex quick release but I do not have any handy to test. I modified the spacing of simple terminal blocks to serve the purpose. Peef most likely chose hard wiring as it will offer better connectivity over the long run.

.. dB

 

I should have posted last night because then I would not have seen @Beefy 's incredible work earlier today

This is the wired up state of the JFET phono - wishing I had a decent signal generator around about now as I would love to see how it measures but I will find some time to plug it into the system and give it an auditory whirl.







..dB

EDIT:

fired it up

it was picking up some high frequency noise from the TT power supply which dropped to inaudible levels when I started the platter turning - definitely need to get it on a scope ( especially with me using an smps )



it sounds very good with the J113

I had recently picked up a soundsmith cart ( SMMC3 from @tommytakis which stemmed from @gaspasser = we are an incestuous lot) to try out something other than MC and have not had much time to listen to the cart so difficult to isolate what is preamp and what is cart but things are very crisp and "slamming" listening to Bonamasa's British Blues Explosion



this phono is well worth building - thanks @peef for sharing your inspired design and layout to make it all easy

 

You give me too much credit.

Awesome build, thank you for sharing your progress and photos!

The RF noise can be really tricky to deal with. I've run into the same issue in some, but not all setups, using the same gear, using either switchers or the Talema toroids in some setups. My sense is that there's some RF junk coupling across the power unit's parasitic capacitance, as I've had some success with wrapping the live wire around a large ferrite before connecting it to the SMPS, or using a proper IEC filter.

Let us know if you try out the other JFETs.

 

Brief update.

1. Noise. The preamp is dead quiet. I grounded the inputs and ran it into my headphone amp at full volume and no audible noise from the pre.

Plugging the TT cables into it brought the return of the noise. I added some strategic ferrite clamps and the noise dropped to below an audible noise floor.

2. The little adapter boards work a treat. Pretty short work to switch between them. The smd options sound obviously "better" /different = seem to extract more definition and air, the CHP3910 is what I have settled on.

Once this has settled in a bit, I will compare it to my other preamps but it is a clear "keeper"

.. dB

 

@peef

Would a 48v laptop power brick work for this?

 

that is pretty much what I have used - 48V SMPS but mine is not as convenient as a laptop supply and generates a little more unnecessary current

 

It bears mentioning that you can mix and match JFETs to voice the preamp to your preference. 2x CHP3910 and 2x 2SK3557 for instance. So long as there is enough current to feed the pre, and leave some current for the shunt regulator (keeping in mind that the SMD parts eat more current). It's prudent to keep a good 10mA in the shunt leg.

Since you socketed the source diodes, you could also drop in a SiC diode or LED to bias the source a bit higher. This reduces current in the amplification stage, but improves overload headroom.

And glad the ferrite trick works! The cascode front end has terrific bandwidth, so it can be picky about what you plug into it. For better RF immunity, the gate stoppers can be increased (R10, etc.) at the penalty of higher noise.

You would need to make some small changes, but it works with as little as 24V. The trick is that the shunt regulator CCS needs to see at least 4V across it or its impedance (and thus the regulator's PSRR) drops significantly.

For a 48V-on-the-dot supply, I would consider replacing Z1-4 with 10V zeners, which should result in a 41-42V regulated supply. The easiest spot to measure the voltage drop across the shunt CCS is from the Vin connector at the top right of the board to the tab of either M3 or M3. For maximum headroom, the cascode's output should sit around half of the regulated supply voltage. There's a test point (labeled "tp) to measure this above the row of caps. This can be fine-tuned by adjusting the value of Z5.

 

thanks for mentioning this, you answered a niggle I had in the back of my mind.

the other niggle = what is the minimum number of jfets in the cascode ?

..dB

 

Just 1. Gain goes up and noise goes down as you add JFETs, as the transconductance of each JFET adds up. One CPH3910 is roughly equivalent to four J113s, in terms of gm/gain.

Edit: And this guy has higher gm, still.

Theory predicts that more JFETs will sound better, but this is audio. Better make sure theory agrees with your ears, it's basically free.

 

In theory, theory and practice are the same.

 

For a lark. I tried one of each type and it sounds very pleasing. Not sure how different it sounds from any other combination but that speaks to the quality of this phono pre. Very pleased with the quality of audio being reproduced..dB