I'm building a JFET preamp with a Transformer based volume control at the output. I'm not saying this is a "bride" to Nelson's beast amp, but it is clearly inspired by his design. As always, many thanks to Pass for sharing these great DIY designs/ideas! I think this might turn out quite good, if I can get the transformer right.
Here is the modeled response of this preamp. Doesn't look like it can provide the 32 mA of current needed below 100Hz. The transformer primary is 100mH and it will have a switch to provide an optional +6 dB of gain. To get 1mH takes 75 turns. So, 100mH is 7500 turns. And the secondary requires 14.500 turns! Threading that much wire through a toroid is going to be a real pain.
Might need more turns or more JFETs for this. I don't want a high output impedance. The JFETs are LSK170C but I plan on using BF862 JFETS.
Primary current with 1 volt peak applied:
Frequency response at output of transformer:
I don't think that the simulated response is really accurate though. Shouldn't it have more bandwidth? The toroid is powdered iron #26 mix with a bandwidth from DC to 500kHz.
I don't know much about transformer design so any helpful comments are welcome!
I might consider an EI type transformer core if I can find a good one. As I understand those transformers should be laminated to reduce eddy current loss and distortion. But, aren't toroids better, but the more difficult way to go?
Here is the modeled response of this preamp. Doesn't look like it can provide the 32 mA of current needed below 100Hz. The transformer primary is 100mH and it will have a switch to provide an optional +6 dB of gain. To get 1mH takes 75 turns. So, 100mH is 7500 turns. And the secondary requires 14.500 turns! Threading that much wire through a toroid is going to be a real pain.
Might need more turns or more JFETs for this. I don't want a high output impedance. The JFETs are LSK170C but I plan on using BF862 JFETS.
Primary current with 1 volt peak applied:
Frequency response at output of transformer:
I don't think that the simulated response is really accurate though. Shouldn't it have more bandwidth? The toroid is powdered iron #26 mix with a bandwidth from DC to 500kHz.
I don't know much about transformer design so any helpful comments are welcome!
I might consider an EI type transformer core if I can find a good one. As I understand those transformers should be laminated to reduce eddy current loss and distortion. But, aren't toroids better, but the more difficult way to go?
Attachments
Last edited:
Hey John,
Don´t think this is the best of ideas. At 20Hz your 100mH has a reactance of 13ohm! Place a descent transformer at the input instead. Probably a 1:2 transformer with at least a primary inductance of 100H would do. You can the dump all except one to three pairs of FETs and add a Slagle AVC at the output.
If going the "transformer after" route I´d advise you go for a commercial TVC like Silk. Think it can take 6V and also has a +6dB winding.
Don´t think this is the best of ideas. At 20Hz your 100mH has a reactance of 13ohm! Place a descent transformer at the input instead. Probably a 1:2 transformer with at least a primary inductance of 100H would do. You can the dump all except one to three pairs of FETs and add a Slagle AVC at the output.
If going the "transformer after" route I´d advise you go for a commercial TVC like Silk. Think it can take 6V and also has a +6dB winding.
Okay, thanks, that would work. But I'm kind of set on doing my own transformer. 10 JFETs in parallel is fine with me. And, I'll do some experiments and raise the impedance of the transformer. Air core transformers are more linear ... maybe I'll just wind a bunch of wire on a spool to get a higher primary inductance. Take more wire but it'd be easier than treading a toroid. The coupling coefficient for air core is like 0.4 I think, so would mean a higher impedance at output, but still less than a commercial volume control I think.
Any concerns about that?
If that's also a bad idea, I may put a good step up transformer at the input, like you mention, and put a 1k pot at the output for volume control. I would only need the gain for LP playback, so I would have a switch to bypass the transformer.
Any concerns about that?
If that's also a bad idea, I may put a good step up transformer at the input, like you mention, and put a 1k pot at the output for volume control. I would only need the gain for LP playback, so I would have a switch to bypass the transformer.
You mean build a new phono preamp. That could be an option since I plan on upgrading my phono cartridge to MC which means a new phono preamp. I'm planning on doing my own.
I'm going to wind some wire on a spool today or tomorrow, and see what I'm in for. Then maybe start looking into phono preamp options.
I'm going to wind some wire on a spool today or tomorrow, and see what I'm in for. Then maybe start looking into phono preamp options.
Hi, I found it flattened the high frequency response and extended it. Don't know why. Still, I don't really trust the spice model of the transformer. I haven't really looked into the generic model too much. I tried to edit the process model in Spice and got an error message saying, you can't do that. I'll do some real world experiments once I roll a transformer see what happens.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.