Second attemp at a hybrid amp.

[Severino]

Hi everyone.

After much indecision and not much time I've decided to finally re-build a guitar amp.
This is the second attemp, the first one worked but the output stage was damaged and I got a motorboating problem.

Ok, let's start with what I built. I had several transformers from old TV's which used tubes, so I decided to use it to make a tube preamp based on a fender preamp, the blackface one which I don't rermember its name, it was pretty easy, Two stages, the volume and the EQ after the first gain stage and a second gain stage that worked as a buffer and as an amp from the passive vol-eq chain. I used an old schematic from about 10 years ago, I think it was a bassman and It required 360V in order to work, so I used a smaller trafo to boost the 210V that the TV transformer gave, using three differents power supplies (with SS diodes and cap filters) I ended with around 350V or so, I plugged it and after a few fixes, (I used a DC supply for the filaments from the same trafo) It worked. The power amp was a mosfet old design and I used a toroidal for this, a PCB and quite a different story. I used the amp connected to a 4 ohm load speakers (8+8 in parallel) so I guess the mosfet were stressed and finally died. Ok, that's ancient history.




After reading about improving the Power amp adding two more mosfet I was happy, I still haven't soldered all parts, but that's not my question (yet).

My question is about biasing the preamp tube, I used a 12ax7/ECC83, looking at some graphs provided in the specification, the old schematic gave arout 360v using a 100K and a 1,5K resistors, my question is if bringing those resistor values down and giving the same mA in "DC-mode" the stages will work as needed, I remember I calculated something about 82K for cathode and 1,2K for anode, I might be wrong, but my question is if there's something more going on when using higher voltages like noise or distortion, remember this is a clean channel. Am I missing something? Or everything?




Thanks, this is my firs issue of many, any help on how to approach the preamp design will be very helpful.

[Severino]

My question is about biasing the preamp tube, I used a 12ax7/ECC83,

Sorry, I was using a ecc82 datasheet which provided voltage and resistors values, here is my shematic, the tone stack may be wrong I only copied the last schematics, I have it stored in some windows-based program and I've used kicad to draw it raw so you could point my mistakes, or maybe refer me to some literature, I have two books from Kuehnel and did all the design using tha theory provided there.


My question is still if the biasing is correct, I did the math several years ago and you could help me if there is some basic mistake.

Kind regards.

[Steve Conner]

The values look reasonable. Measure the plate voltage at idle. If the tube is biased properly, it should be about two-thirds of the supply voltage. I'd say a range of one-half to three-quarters is acceptable, but outside that, it'll sound crappy when overdriven.

The last amp I built was a hybrid too.

[Severino]

I'm still looking for the notebook where I did the math, I still haven't build it, I have all the components, though.
I do not want to overdrive these stages, I don't know if you mean that It will sound crappy when passing an overdriven signal or when actually overdrive the second stage. I'm still remembering my student years with transistors and calculating the bias point so the clipping will happen simmetrically. (Excuse me for my poor translation to english).
I actually don't remember the name for this but was to provide maximum range of amplification and put the signal in the center of a drawing we made to calculate the Dc-point.
Ok, I'm making a mess about this.
I will try this way: When you mean crappy is it about the bias point not being right and clipping being not symmetrical? Or do the "low-voltage" design have something to do with it?

[Severino]

The values look reasonable. Measure the plate voltage at idle. If the tube is biased properly, it should be about two-thirds of the supply voltage. I'd say a range of one-half to three-quarters is acceptable, but outside that, it'll sound crappy when overdriven.

I simulated with a program and gave 82V on the plate at a 196V of B+. I changed the tone circuit for the 5F6-A, but I did not changed the structure of the preamp.



82/196 = 0,42, a 42% of the power supply is the voltage on the plate.
I also noticed that when clipping it's not symetrical, so bias point is wong. I'll do the math again.

Thanks.

[Steve Conner]

Yes, I meant that it'll sound bad if the clipping is very asymmetrical.

I once tried adding adjustable bias to the second preamp tube of an amp as an experiment. It sounded best when biased symmetrically for maximum unclipped output.

[Severino]

You are right, using a simulation software I've found that it wasn't biased right. After taking an overlook at "guitar amp preamps" there's something wrong with my math or maybe with my approach at the preamp. My math said 3mA for an ecc83, but using simulation is only 1,2mA which brings the DC point way low on the curves supplied with the tubes.

Since I'm looking for a simple input stage with eq I shouldn't use high gain triodes, I keep lowering the amplitude of the signal via voltage dividers and something tells me that gain here is not the issue. Finally (on simulation) I used a 68k plate resistor and a 680ohms cathode resistor and the plate voltage was around 70% of the power supply, the clipping wasn't exactly symmetrical, but close enough.
The problem that I see now, besides the gain, is the eq network. I will try to found the transfer function of some vintage eq-stage and change the values to fit a much lower gain approach.
Do you think ECC81 could do the trick? I will use two and use cathode followers to isolate the gain from the eq, and the eq from the "master", and maybe a mosfet to isolate the preamp from the amp, just ideas I will try to put into numbers next week.
I've seen biasing techniques on radiotron's which are not fixed and maybe that's a nice way to go on all stages. Thanks.

[Steve Conner]

My personal opinion is, to get the best sounding tube distortion you have to use ECC83s, and at least 200V supply voltage, preferably 250. I've tried the 81 and 82 but always ended up going back to the 83.

It's a lot easier to arrange a 250V supply in a hybrid amp and copy a classic Fender preamp, than it is to redesign the classic tube amp architecture for the same dynamic "feel" with a lower supply voltage, which leads to a different gain structure.

I always try to include a transformer in the signal path of my hybrids. I find it makes a big difference. For my last one, I used an EL84 as a driver and a Valve Jr. output transformer as an interstage transformer, to drive a transistor output stage. This gave a big improvement compared to just coupling a tube preamp directly to a SS power amp.

[Severino]

My personal opinion is, to get the best sounding tube distortion you have to use ECC83s, and at least 200V supply voltage, preferably 250. I've tried the 81 and 82 but always ended up going back to the 83.

It's a lot easier to arrange a 250V supply in a hybrid amp and copy a classic Fender preamp, than it is to redesign the classic tube amp architecture for the same dynamic "feel" with a lower supply voltage, which leads to a different gain structure.

I always try to include a transformer in the signal path of my hybrids. I find it makes a big difference. For my last one, I used an EL84 as a driver and a Valve Jr. output transformer as an interstage transformer, to drive a transistor output stage. This gave a big improvement compared to just coupling a tube preamp directly to a SS power amp.

You're right, If you want distortion on the amp, ecc83 preamp and EL84 pow-amp is the way to go.
My hybrid "ideal" amp did have a el84 in class A with an output transformer, but right now that's not my goal.

I coupled the preamp (a blackface/boogie preamp) directly with the SS PA and It sounded great. Before running this mosfet PA I built the preamp and used it on the FX-loop in of a vintage peavey and the difference was huge, the sound was great.

I did not overdrive the preamp, I had then a dist pedal and again the difference was big. So, with nowadays DSP's on multi-effects and simulations, etc, my goal is a simple tube preamp with EQ, and I don't want to overdrive it. I'll keep the signal processing before the preamp.

I've found, as I said that 68k as plate resistor and 680 on the cathode with a 22uF electrolitic keeps the gain high enough to drive the EQ network, and pass the second amplification stage without clipping with a single ECC83 tube. I wil install a SS buffer to isolate the output of the second stage from the input of the PA which is line-level. I can, however install another transformer to get to 340V and then have the actual voltages of the clone preamp. Anyway, I don't want do use another tube in the chain, so I will google for a buffer stage. I will make the complete schematic and will post it here. I must start soldering and listening soon.

[Steve Conner]

Are you looking for a buffer to drive the power amp? I'd use an op-amp buffer after the master volume. A low-distortion FET input op-amp like the OPA134.

You could also use a discrete JFET. The idea is to put it directly at the PA input, because then you don't need a high voltage buffer.

The LND150 MOSFET can be used as a high voltage source follower.

[Severino]

I always try to include a transformer in the signal path of my hybrids. I find it makes a big difference. For my last one, I used an EL84 as a driver and a Valve Jr. output transformer as an interstage transformer, to drive a transistor output stage. This gave a big improvement compared to just coupling a tube preamp directly to a SS power amp.

Hi. I've found on the corner of my workshop a little audio transformer 5,6k->4ohms, I think by the size of it was used on a TV speaker, so, not much power, but enough to really isolate the power stage and the preamp.

Here is a schematic of the PA, It's built and maybe that's not the idea of a hybrid that you have in mind.


About the preamp I'm still struggling to find the power transformer between a lot that I have around, so maybe I will have something done by the end of the week, at leasto to plug into the PA of another amp and listen.

A jfet seems nice, but couplig using this little trafo could be the way to go. Any useful thread o website to found information about that EL84 coupling?

The input impedance of the PA is around 68K, but can be modified to a lower value if needed. The input capacitor can be removed if the coupling is through trafo.

Ideas? Links to information about this?

Regards.

[loudthud]

One thing you must avoid is driving the solid state power amp with too much voltage. It's ok to drive the output to clipping, but too much input beyond that will kill the input transistors or cut off the current mirror Q5, Q6 and thus Q7 causing the output to slam the negative rail when it should be on the positive rail. A zener clamp works well to do this. Something like 6V to 12V diodes will work fine.

[Severino]

I'm lost and sleep-deprived.

One thing you must avoid is driving the solid state power amp with too much voltage. It's ok to drive the output to clipping, but too much input beyond that will kill the input transistors or cut off the current mirror Q5, Q6 and thus Q7 causing the output to slam the negative rail when it should be on the positive rail. A zener clamp works well to do this. Something like 6V to 12V diodes will work fine.

How and where should I put the zeners?

I'm still looking for a design to isolate the preamp using a transformer much like the spring-reverb tank driving circuit, I've seen some designs using el84 's but the output keeps being too high for the 2Vpp line-level that is expected on the PA input.

I'll keep reading, but anyone can point in the right direction?

Right now my design is as follows:

Input - 12ax7 amp stage1 - EQ network - Gain (tuned not to overdrive the following stage in worse-case scenario) - 12ax7 amp stage 2 - Master volume - buffer - PA - 8 ohms load (4xV30).

I'm stuck in the buffer, a cathode follower using a 12ax7 will leave a stage of the tube useless, a EL84 pentode "reverb-style" is the best I can think to give another feeling to the amp, but... beyond that I must say I have to get my head on a specific design (for the interrface) and start from there. Any ideas?

Again I'm sorry if I can't make myself clearer because of my english.

Kind regards.

[loudthud]

Without knowing the specific component values on the power amp schematic, I suggest something like this:

[Steve Conner]

You want to use an output transformer? Just build a normal tube amp output stage and load the OPT secondary with a dummy load resistor. Connect a 10k pot in parallel with the resistor, this is your "master volume". Feed the power amp input from the wiper. If the output voltage is too much, don't turn the pot up all the way, or add some resistance in series with the clockwise end. But 3w into 4 ohms is only a few volts so this may not be needed.

One advantage of this approach is, you can rig up a "low power" switch that disconnects the SS PA and drives the speaker directly from the small tube amp. Great for practice at home.