Liverpool Circuit - Princeton Chassis - 7 pin output tubes

[Cliff Schecht]

I went from Teflon wire to PVC wire in an express build and the change in tone was big enough to be worth the effort, just made the amp sound less aggressive in the top end.

It would have been interesting to do that incrementally, one wire at a time, starting at the input. If there is any difference, I would expect it at the longer, higher voltage leads, early in the chain. Since I correlate the material difference to lead-to-chassis capacitance, the difference would be minimized by lifting both types of wire away from the chassis.

If it is a capacitance effect, a few prudently placed picofarad caps should be able to similarly tame the harshness. Snubber caps - ala the Dumble HRMs - come to mind.

That is also the one effect due to solid core wire - being able to lay the wire down against the chassis and keep it there.

Or it could be the color...

P.S. You're hanging it out there, RJ. I think you're going to have some cool results with this. I just hope you get it done before the Wreck Police come back from vacation.

I think a better approach would be to stick the amp on a network analyzer and take screenshots as you chopstick around in the amp or change wires/add snubbers.

[Cliff Schecht]

I went from Teflon wire to PVC wire in an express build and the change in tone was big enough to be worth the effort, just made the amp sound less aggressive in the top end.

It would have been interesting to do that incrementally, one wire at a time, starting at the input. If there is any difference, I would expect it at the longer, higher voltage leads, early in the chain. Since I correlate the material difference to lead-to-chassis capacitance, the difference would be minimized by lifting both types of wire away from the chassis.

If it is a capacitance effect, a few prudently placed picofarad caps should be able to similarly tame the harshness. Snubber caps - ala the Dumble HRMs - come to mind.

That is also the one effect due to solid core wire - being able to lay the wire down against the chassis and keep it there.

Or it could be the color...

P.S. You're hanging it out there, RJ. I think you're going to have some cool results with this. I just hope you get it done before the Wreck Police come back from vacation.

I think a better approach would be to stick the amp on a network analyzer and take screenshots as you chopstick around in the amp or change wires/add snubbers.

[RJ Guitars]

...I think a better approach would be to stick the amp on a network analyzer and take screenshots as you chopstick around in the amp or change wires/add snubbers.

Having access to a network analyzer would be very cool, but that isn't on the Christmas list at this year. Allyn Meyers posted something about a spectrum analyzer once. That seems like the ultimate piece of technical equipment for evaluating and comparing amplifiers... again that isn't likely to be under the tree this year either.

I do have an oscilloscope and I can see the value of getting a visual of the before and after. As far as using chopsticks it would be nice to eliminate the need for that technique by using actual capacitors where needed. In my current design I could use PVC insulated wires but that is potentially a step backwards. Finding a way to understand and duplicate the capacitive coupling of the preamp wires is really the big driver in this effort. I also expect there are some other factors to be sorted out in the "edge of stability" behavior of the Trainwreck amp designs.

I don't know any tube amp builders with network or spectrum analyzers but I'd be curious how common that is.

rj

[Cliff Schecht]

I have and/or have access to all of the necessary equipment.. I might try to rig up this test over the break.

[burkeee]

I have some experience using tiny output tubes. There are enough unknowns using out of the ordinary tubes that there are plenty of optimizations to do before you get into finicky things like teflon vs PVC wire. If there is little clean headroom there is probably too much voltage swing from the preamp for those little guys. You could try using split plate resistors to cut the gain without changing the way the circuit acts. Play with the bias for the output tubes. I've found that biasing them a little colder than datasheet values can cut some harshness.

I also find myself running small watt amps flat out, and sometimes circuits from big tube amps that we might typically use at 1/2-3/4 volume just sound harsh flat out.

[RJ Guitars]

I have and/or have access to all of the necessary equipment.. I might try to rig up this test over the break.

Cliff, If you have the tools, time, and inclination to really look at this and publish the results, I'd be willing to send you a chassis with some tube sockets and some wire including some of "the wire" that is rumored to have been used in the real Wrecks as well as the finest teflon hookup wire I have.

If we could actually get a look at the capacitive behavior of the wire and the chassis, and eventually get video spectrums on a finished amplifier, that would be very cool stuff. That should ultimately help us sort out an "equivalent circuit" and give us a handful of taming tools for forum builders. I'd be quite pleased to contribute to the effort.

Let me know if you you have any interest in the idea.

Thanks,

rj

[RJ Guitars]

I have some experience using tiny output tubes. There are enough unknowns using out of the ordinary tubes that there are plenty of optimizations to do before you get into finicky things like teflon vs PVC wire. If there is little clean headroom there is probably too much voltage swing from the preamp for those little guys. You could try using split plate resistors to cut the gain without changing the way the circuit acts. Play with the bias for the output tubes. I've found that biasing them a little colder than datasheet values can cut some harshness.

I also find myself running small watt amps flat out, and sometimes circuits from big tube amps that we might typically use at 1/2-3/4 volume just sound harsh flat out.

Good stuff and it would be beneficial for all of us if you'd be willing to share a bit more of the technical details of what you know.

In this case the output tubes are small, but not tiny. The 6AQ5 is said to be electronically the same as the 6V6GT tube, only the packaging and voltage ratings are reduced. However I am in agreement that your thoughts about mucho voltage swing are on the right track and some similar ideas to what you have mentioned are planned to reduce the signal entry level into the power section. The "PIG" (Phase Inverter Gain) control will likely be the first implementation to balance the the input with the handling capability of the power amp tubes.

The preamp tubes are 12AX7's (at least for now) and the harshness in using teflon wire in the preamp section of a Trainwreck is a previously established fact with these amps. I don't expect anything different in this build design.

In considering the wire and the potential for harshness to exist due to the use of teflon wire it's important to note that wire is a passive contributor to the circuit. That is, wire does not add the harshness, it only lets something that exists get through to the output. Where that harshness initiates isn't so important to me, but removing it certainly is. We've all read about Geetar Pickers use of a 40' feet of input cable to take the edge off off of things so he can use the bright switch. Same story again, the cable doesn't technically add smoothness, but something electronically significant is happening. Can it be measured, quantified, and duplicated?

rj

[burkeee]

Looking at the datasheet, the 6BQ5 at 250V is similar to the 6V6 with 250V on the plates, but pretty much every amp used the 6V6 with much more juice on the plates. The voltage swing for max power at 250V for 6BQ5 or 6V6 is only ~30V, but a typical 6V6 guitar amp might take a 50V swing. So the preamp output and biasing would be different from a typical 6V6 design. But the liverpool is an EL84 amp right? EL84s only need ~15V for full power. Are you using the liverpool circuit verbatim, or ??

I'm no trainwrech or Ken Fischer expert, so I looked up his comments on teflon vs pvc wire. His point was that silver coated copper wire leads to the harshness, and said that teflon wire is silver coated. But it seems to me a lot of good quality wire we'd use for the temp and voltage range of a tube amp would be silver coated, regardless of whether it was teflon or pvc.

[RJ Guitars]

Looking at the datasheet, the 6BQ5 at 250V is similar to the 6V6 with 250V on the plates, but pretty much every amp used the 6V6 with much more juice on the plates. The voltage swing for max power at 250V for 6BQ5 or 6V6 is only ~30V, but a typical 6V6 guitar amp might take a 50V swing. So the preamp output and biasing would be different from a typical 6V6 design. But the liverpool is an EL84 amp right? EL84s only need ~15V for full power. Are you using the liverpool circuit verbatim, or ??

I'm no trainwrech or Ken Fischer expert, so I looked up his comments on teflon vs pvc wire. His point was that silver coated copper wire leads to the harshness, and said that teflon wire is silver coated. But it seems to me a lot of good quality wire we'd use for the temp and voltage range of a tube amp would be silver coated, regardless of whether it was teflon or pvc.

Burkee , thanks for the note. In between major winter storms and moving to a new office I haven't been doing much amp tinkering over the Holidays. However, last night I set about in earnest to make this lightweight Liverpool do it's thing.

I had some noise and an intermittent low output problem that was sucking the joy out of this build. I found a voltage discrepancy between each side of the phase inverter and worked late into the night last night and reflowed the solder on just about every connection on the amp. No good so I changed a tube socket, and then it all seemed to come together. I was kind to the neighbors and waited until today to crank it up and the demon was back. No fun!! I was doing something when I realized one output tube was too hot to pull out and the other was barely warm... I should have sorted that out quicker!

New set of tubes and now it's sounding pretty cool. Way plenty of gain but very cool sounding. It does clean to mean but has much more mean than clean.

So now the circuit and the tubes are healthy and I can start dialing it in. The clean is good but I am used to Rocket amp clean tones and that sets the bar pretty high. I need to do some work to get there.

I played with some output tube options after I had the intermittent tube problem but I'll put that into a different post.

I used the Liverpool schematic verbatum except I now have a 300 ohm cathode bias resistor in there. I've been reading Merlin's book on preamp design and going over the tube load-line stuff. That book has been an eye opener for me in recognizing some portion of what I don't know.

Now I think I can start matching up some of those signals. I'll appreciate any help along the way in setting it up so I can get a little more clean out of it and wouldn't mind smoothing up the grind just a touch as well.

thanks,

rj

[RJ Guitars]

Hey guys,

One of the drivers for me in putting together this undersized Liverpool was to explore output tube options for building a low wattage amp with Trainwreck tone.

After I had a little trouble with one of my output tubes earlier today i got out the entire supply of 7 pin tubes and gave a bunch of them a try. There are a big handfull of tubes with the same pinout as the 6AQ5 plus a bunch more that will work if you jumper pins 2 and 7 together - 6AK6 for one. I've done this before using a DPDT switch.

For todays sampling I used the following:

EL95
6AQ5 / 6005
6AN5
6AK5
9001

The 9001 and 6AK5 are little bitty rascals that I call "acorn" tubes. The 6AK5 sounded better than the 9001 but I think I was running it pretty far above the rated voltage so I didn't mess with that too long. The best sounding one of the lot was actually the 6AN5 tube. As I sort out what is needed to match up the output tubes with the preamp any of these could be a lightweight contender for powering a Wreck.

Fun stuff and plenty to be learned.

rj

[Ian444]

I used the Liverpool schematic verbatum except I now have a 300 ohm cathode bias resistor in there.

Which valve has the 300 ohm cathode resistor?

Now I think I can start matching up some of those signals. I'll appreciate any help along the way in setting it up so I can get a little more clean out of it and wouldn't mind smoothing up the grind just a touch as well.

I guess you were asking for experienced builders' opinions, hope its OK to throw in my 2c worth. I've been mucking around with a Liverpool cct with 2 x 6P1P output tubes, these are similar to 6AQ5 but 9-pin, cost about $2 plus shipping from ebay last time I bought some. I am very new to TW ccts, and fairly new to tube amps. I found a few things that reduce the drive (grind?), the grid leak resistor on pin 2 of V2a can be made smaller, also the grid leak resistors on the output tubes have an effect, and the tail resistor in the PI also has a (lesser) effect. The amount of feedback also seems to affect the amount of cleanness and meanness. I think C8 needs to be as small as possible to keep a tight tone (not too bassy/muddy), but I have run out of caps to experiment with. I need to order a heap.

I am wondering why the Liverpool tone stack uses the 1M treble pot and 50pF cap, and some Express ccts use a 250K treble pot with 500pF cap. I would like to try both and see how they affect the sound.

[ChrisM]

I am wondering why the Liverpool tone stack uses the 1M treble pot and 50pF cap, and some Express ccts use a 250K treble pot with 500pF cap. I would like to try both and see how they affect the sound.

One is a Vox TS the other is a more typical Marshall/Fender TS. Different designs.

[RJ Guitars]

I went to lunch with Zippy today - not only because he is good company but because he has a great technical mind and doesn't charge too much for sharing that knowledge. I described the Liverpool amp behavior, particularly the small amount of headroom before the distortion came on strong.

He talked about phase inverter balance and gave me another tutorial on load lines, bias points, and voltage swing. Some of this will eventually soak in. Then he drew me some pictures showing a distorted sine wave that was flat on one side and normal on the opposite side. He said he believed that if I was to input a sine wave into my amplifier and hook an oscilloscope to the output that I would like see this in real time as I ran the volume control up.

I do have some nice tools and the oscilloscope is one that I probably under utilize. I am usually inclined to just change parts until my ears are happy but this was easy enough and much of this is new territory for me. No sense it making it harder than it has to be so I did the measurements he suggested and I saw exactly what he had described. At very low volume levels a nice looking sine wave. As I crank up the volume the lower side of the wave started to flatten out while the upper side remained normal. At very high volume levels the entire wave form looked pretty crazy... but I'll leave that to sort out later.

I decided to build this using the Liverpool schematic values for the EL84's while actually using the 6AQ5's, not because I expected it to sound good but I wanted to map the process of sorting it out and allow anyone that was curious to learn along with me. I know some of you are not that slow and are way ahead of me. A few folks already indicated that bias point and voltage swing on the phase inverter would be important... I now have a better grasp on how all the pieces fit together.

I'll set out to adjust the bias point and hopefully get some more scope shots. If that goes well then I'll look into the voltage swing after that.

rj

[Cliff Schecht]

I can see the bias shift in that waveform, it looks like crossover distortion. Some assymetrical clipping is nice, but too much sounds flabby and weird. Some things like this you can only see on an oscope, at least at first until you start to learn what the waveforms sound like without having to hear them (or vice versa, knowing what waves look like just by hearing them). Here's what I said earlier today over at TDPRI about scopes:

Yeah I use my scope constantly for amp debugging. They make it easy to trace out signal paths, look at waveforms at various points and you can use them to determine gain as said above. On nicer scopes you can also get automatic readouts for peak-to-peak, peak and RMS voltage, phase information, rise and fall times, pulse width, delay, frequency and FFT's to look at your harmonic content. Most of these measurement features are available even on good older Tektroniks (90's models) and most new digital scopes over a few hundred dollars (probably closer to $1000 for a decent scope though).

There are a couple of things to keep in mind when looking at a scopes for tube amps though. What is the maximum voltage the input section can handle? How many channels is it? What is the speed and available time bases (determines what frequencies you can see). If you are looking at square waves, a 1 MHz bandwidth is equivalent to only a 1uS rise time which is not necessarily fast enough to see clean square waves. Worst case scenario here is if you work with switching power supplies and such with nS rise times, you need a scope that can go into the GHz range. Also if your amplifier is oscillating in the MHz range, you won't be able to see that with only a 1 MHz bandwidth.

Something that you will learn to do after hours and hours of staring at a scope is be able to look at a waveform and know where the harmonics are at. This is an incredibly important skill that helps not only SEE what the amp sounds like but it can also be a powerful troubleshooting/amp voicing skill. I can look at a distorted guitar amp waveform and tell you what characteristics the amplifier will have based on just waveshape. For example if you see a lot of rounding on the rising and falling edges of a square wave, this means you have excess lowpass filtering going on. If you have a waveform feeding into a preamp tube grid that is causing it to shift the bias at full-on distortion then you have a blocking distortion problem. Lots of little issues become clearly visible on a scope that would otherwise leave you scratching your head. Just know the limitations and be careful not to shove 500V on a 100V device (I put over 1000V on a Fluke and blew it recently actually hehe).

As far as getting down biasing, it's all a matter of PRACTICE. You need to practice going from paper, building, and measuring your results and comparing them to your calculated results (you did calculate your gain and bandwidth, right? ). Lots of great old texts on the subject but I always go to my big red RCA tube bible. Get yourself a copy of RCA Electron Tube Design (I think that's the title) if you don't have one (you can download it now too). I also reference Merlin's site a lot for quick reminders and such, he's probably my other most often visited site. The Adam's triode calculator is quite useful for quick calculations and comparisons of different 12A*7 tube gains.

Simulation is also a powerful tool. It's only as smart as you are though, so you have to do a bit of work to figure out how to emulate things like transformer parasitics, amplifier parasitics and reactive speaker loads. I use simulation a lot when working on new designs, it's especially useful for figuring out your preamp bandwidth and gain figures. The power amp goofiness is harder to emulate but with practice it can be achieved. I used to use Altium Designer for simulation but recently discovered NI's Multisim has a lot of great models built in and let's you do a lot of real-time analysis (i.e. they have a built in scope, function gen, spectrum and network analyzer that you can tweak in real time, quite useful!).

[Tillydog]

I decided to build this using the Liverpool schematic values for the EL84's while actually using the 6AQ5's, not because I expected it to sound good but I wanted to map the process of sorting it out and allow anyone that was curious to learn along with me.

This is probably heresy, but is offered FWIW:

I put a pair of EL84s on an Express preamp - it started as an AX84 4-4-0 with a 12AY7 as the PI. I found that while the 12AY7 reduced the PI gain, it increased the available voltage swing from the PI. This became a problem because it allowed the negative feedback to put huge -ve spikes on the PI output once the the output grids were starting to conduct, shifting the drive signal up.

Cutting a long story short, I went back to a 12AX7 PI, but with ~ 40% split load resistors (47k/33k ILO 82k & 56k/39k ILO 100k - lowest value resistor to B+) & 100K grid leak resistors on the EL84s. This allowed the PI to reach full output just after the EL84s start to over-drive, and limits the extent of the NFB 'nipple' on the bottom of the PI output.

I tried a 35% split load ratio first, and while this was great for headroom, it ran into muddy distortion almost immediately. The 40% ratio sounds fine to me. The resistor values reflect what was at hand - refining these might pay dividends.

Also putting a whopping cathode bypass cap on the EL84s (2200uF) smoothed things out a surprising amount, and added a bit more headroom.

No idea if this is right or wrong, but I'm happier with the amp now.

As ever, YMMV

Andy
(Long time lurker)