Low Impedance output for B15n style preamp

[Martin Beer]

I'm planning to build a simple preamp for bass, based on the Ampeg B15n, which I want to couple with a class-D power amp with 10kohm input impedance. I intend to copy the first two triodes and tonestack of the B15n (i.e up to the ext. amp jack in the schematic), using a 6SL7. 10Kohms seems a bit low to run off the plate of this preamp, so I'm anticipating adding some form of buffer to the output. Would there be any tonal advantage to adding a second valve to build a cathode follower on the end, or would I better off looking at a MOSFET source follower as per the Geofex MOSFET follies article? As I'm using a solid state power amp, I'm not necessarily averse to having a transistor in there, but I'd need to do a bit of reading to do it properly, as I haven't looked at high-voltage MOSFETS before. I guess another option would be to have a transformer output, something like a microphone matching transformer in reverse. I'd be interested to see which approach people would recommend.

[soundguruman]

Using a microphone is always preferable. A cathode follower is OK. There are all different methods, but none are as good as a microphone. After all, the microphone is the most isolation and the highest frequency response, and it sounds 'real.'
Line out really doesn't sound real.
I mean ideally, for hum and noise you can use a transformer. That's the quietest. BUT it's a lot of money for a transformer with 'hi fi' frequency response. Hi fi is what you want for Bass. The cheaper smaller transformers have suckey frequency response.
So the mic being cheaper and just as hi fi...and you are getting the compression of the output amp and speaker.
Transistors are heresy. ha ha ha he he he hoh.
Yes there are balanced line drivers IC chips, all in one for audio. Maybe that would be cheap and good to drive a PA board, or mixer, etc...
but it doesn't sound like tubes anymore, ya know?

[Martin Beer]

I can see where you're coming from, but I'm not trying to capture the whole sound of a B15n - I'm just looking to build a nice musical sounding preamp, and I like James style tone controls, so the B15n pre seems a good candidate. I want it to do a similar job to something like an Alembic F2B but with a different flavour and without the midrange scoop. So it seems like a reasonable approach to build just a preamp and add some form of buffer to make the output more usable with solid state power amps or into a desk.

[soundguruman]

I can see where you're coming from, but I'm not trying to capture the whole sound of a B15n - I'm just looking to build a nice musical sounding preamp, and I like James style tone controls, so the B15n pre seems a good candidate. I want it to do a similar job to something like an Alembic F2B but with a different flavour and without the midrange scoop. So it seems like a reasonable approach to build just a preamp and add some form of buffer to make the output more usable with solid state power amps or into a desk.

you want to put transistors in a tube amp?
just buy a PV or a music man instead.

[Martin Beer]

OK, I get it, you're a tube purist. Consider your point to have been made. Just to clarify in case anyone else wishes to comment, I'm not talking about modifying an existing amplifier, I'm talking about building a preamp from scratch, to pair with a SS power amp. I was just hoping for some input and opinions on what type of output buffer might be worth trying for my application.

[JoeM]

Martin - I think your idea of a buffer of some sort is fine. Just because you've got some kind of SS device doesn't mean it's going to ruin the sound. There are several forum members with the technical chops to help you, hopefully they'll pitch in.

[Martin Beer]

At the moment I'm looking at either using a MOSFET buffer using an LND150 like some of the FX send schematics I've seen, or using another valve (possibly another octal like a 6SN7) to add an AC coupled cathode follower. I'm curious to hear any views on the relative merits of either approach, and whether any other types of output stage would be useful for my application, which will be fairly clean bass guitar.

[Steve Conner]

Last time I did this, I used an op-amp buffer. My Class-D module had +/-12V aux supplies, so it was an easy choice.

You can easily arrange the gain structure so the buffer never clips. As an example, I used a diode limiter before the buffer to soft-clip everything to about 8V p-p, because my Class-D amp made some really nasty noises if it was overdriven too heavily. You may sneer, but this same trick was used in the original SVT to save the power amp from really excessive drive.

Another advantage: My power amp had a balanced input. By using a dual op-amp it was easy to arrange balanced drive, though it arguably wasn't needed for such a short connection. However, every bit of interference rejection helps. I had trouble with switching noise from the output stage getting back into the preamp input, causing some really horrible-sounding distortion.

The project was going great until the Class-D module underwent a pyrotechnic self-disassembly.

[Martin Beer]

That's not a bad idea, as I will have an auxilliary supply on board. Would your diode limiter be as simple as a couple of Zeners back to back, or something more sophisticated? I'd be curious to know which module you were using. I'm looking at Hypex and Coldamp, as they seem to offer about the best support and documentation for DIYers. The B&O Icepower modules look good, but not easily available.

[Steve Conner]

Hi Martin

I wouldn't recommend Coldamp: I had a disastrous experience with them. I bought a BP4078 and SPS400 from them, and the power supply exploded twice and the amp module once.

After some heated debate between myself, Coldamp, and a friend who designs SMPS for a living, Coldamp's designer admitted that he didn't own a variac and had never tested the SMPS above 220V line voltage. For highly technical reasons that I can't go into here, it was disastrously unstable above about 235V. My line voltage was reaching 250V at some times of day, and that made it blow its guts.

The BP4078 amp module died later, when I forgot to turn the cooling fan on during a dummy load test run. It put the full +65V DC to the load, which would have destroyed any speakers instantly. An autopsy revealed one shorted MOSFET, but the heatsink surface underneath it was uneven with a large burr around the mounting hole, which surely contributed to its death.

Lately, I got both the SMPS and the amp module working again. I modified the SMPS to be stable at 250V line, and rebuilt the amp with 4 MOSFETs instead of 2, and better thermal washers and clamping. However, I'm not sure I trust it to run without a speaker protection relay now, and that takes up yet more space that I was hoping to save. Maybe I'll try the old Peavey triac clamp circuit.

To cut a long story short, I would surely buy Hypex or Icepower next time. I've also heard good reports on the units from ClassDAudio.com: Class D Audio Home page

If Sergio is reading this, sorry, but it's the truth! You can send me a replacement as long as it's not made by Coldamp! To be fair, this happened many years ago and their products may have improved. The new SPS400 looks different.

My limiter was a high voltage diode from a microwave oven, inside a bridge of 1N4148s. I like the soft clipping characteristic of the HV diode, and its huge forward voltage drop is nice too (it is about 10 diodes in series internally) The diode bridge clipper in Marshall's JCM900 amps works the same: it uses a high voltage bridge rectifier for softer clipping than Zeners or signal diodes.

[J M Fahey]

Agree and add: I would add a suitable attenuator (that preamp oputput is *hot*) plus *any* buffer: FET, MOS, bipolar or Op Amp.
Unity gain buffers usuall change nothing of the signal going through them, and then you are driving an SS amp.
By the way, since I've thrown compactness , low weight, etc. overboard, why use a Digital amp at all?
What's wrong with a simple chipamp or maybe some discrete one for higher power?
I removes one more unknown variable from the equation.
Plus analog SS amps are mature by now (50 y.o.), and we know how they clip and can live with it; Class D amps have not reached that stage yet.

[Martin Beer]

Steve, thanks for the heads-up on Coldamp. I'm not far away from you, so I'd presume the voltage in my town will be similarly variable. The Hypex amp modules at least have the distinction of being used by a couple of bass amp manufacturers, like Hevos, which I hope bodes well for their reliability.
JM, I'd consider using an analogue SS power amp, though I'd have to find a good module or kit as my design skills are otherwise not up to the task. The lightness of the SMPS stuff is attractive though, especially as I live up stairs and can't often park very close.

[olddawg]

People always make things difficult. You can get an Art tube preamp for $29 and it has an unbalanced line out, a balanced line out as well as a voicing selector switch, mic/Line inputs, phantom power, level controls, VU meter, limiting, etc. it makes a great DI box or a mic preamp. Just plug the unbalanced line level out of anything into one. Problem solved.

[Martin Beer]

I did a bit more reading on the Hypex modules. They have 100K ohm input impedance, and 2v sensitivity. I'm thinking that if I use one of these, I might be able to go without a buffer. If I was to pad down the preamp output with a simple resistor divider, the preamp output would never see less than a few hundred K ohms, which seems like an acceptable range. I'd need to experiment a little to find the right values, and incorporate a diode limiter for the peaks as Steve suggested. I'm still not averse to adding a buffer if it proves necessary, but I think I will try it this way first.