Monday, April 24, 2017

How to Make a Pentode Voltage Amplifier That is More Linear Than Any Triode

It is generally accepted that pentodes are on average a good bit less linear as voltage amplifiers than triodes. Today I'm going to show how to take a pentode and make it more linear than any triode, or at least any triode that I have any experience with (if there is one that beats this performance, I wouldn't mind knowing about it).

I've previously written on this blog about highly-linear power stages using KT88s. The approach here will be similar, using voltage feedback that is parallel applied and driven with a p-channel FET, so the overall input impedance of the circuit will be very high and easy to drive. The p-channel FET makes things work out with a minimum of components. There are other variations of this circuit that would be possible with n-channel devices or tube followers/active loads should one desire to keep it all tubes.

Conceptually, this is kind of what I'm attempting to do as far as feedback scheme goes:

And here is a more fleshed-out conceptual drawing of the actual circuit:

What I've done is make a very high gain amplifier by putting an active load on a pentode, then I am using local feedback to reduce my distortion down to minuscule levels. I drive the feedback network from the source of the mosfet in the active load (a low impedance drive point) to keep the feedback network from loading the high-impedance point at the pentode plate. The circuit works very well and is very simple.Feedback-phobia is unwarranted as will be shown in the distortion spectra, which reveals very little distortion and what is there being low-order.

However, before I get too far perhaps I should take a slight detour and explain why I did these experiments with big power pentodes like the EL84 and EL34. I started this experiment trying to find a replacement for the driver stage in my Unity-Coupled amp. It requires 160Vrms to drive the output stage to clipping so it requires a tube capable of idling at 350 or more volts. I also wanted to develop a driver for an SE amp with a follower output stage, which would require a tube that can idle at 500V or more on the plate. EL34 was the cheapest tube I could find that fit the bill. It is overkill but it was the cheapest one. I also had an EL84 laying around so I ran some tests with it. I also had several beam tubes around and tested some of them but the results were not as good as the true pentodes. This makes a bit of sense, since the characteristics of beam tubes get kinky at low plate voltages and low currents. The performance of the beam tubes were still quite good, just not as good as the pentodes.

These concepts could be applied to small signal pentodes as well. I haven't tried it yet, but if/when I get the chance I will.

Here is a simplified version of the practical circuit (I have left out gate protection diodes and gate stoppers):

With these feedback resistors I get a gain of a little over 13. I spent a lot of time optimizing plate voltage and screen voltage and found that the higher the plate idle voltage, the lower the distortion. The opposite was true of screen voltage, the lower the voltage the lower the distortion. Of course, with the screen you can only go so low until you clip at the Vg = 0 line so you have to find that point and back off a little. For the EL34, this was 35V at a 10mA operating point.

I also had a bit of peaking in the frequency response that showed in square wave testing. I solved this with adding 12pF in parallel with the feedback resistor. This also limits bandwidth to 200kHz, which is really plenty. Earlier tests showed the response to be 3dB down at 350kHz. With a 10M90S as the upper device in the active load, there was no peaking in the response but with a 1700V IXYS part, there was peaking and ringing in the square wave before adding the capacitor.

Here are the distortion results I got with the plate set at 600V and 35V on the screen:











10kHz Square Wave

Here are some earlier tests on an EL84 with the plate idle voltage set to 225V and 100V on the screen:







I think that I could have gotten the EL84 distortion down more. I performed the EL84 testing before I discovered that lowering screen voltage reduces distortion. I will revisit that testing when I get a chance.

Friday, March 24, 2017

P-Channel Shunt Feedback KT88 Amp Revisited

This is Part 2 of a series on an amp I designed and built. Click here for Part 1.

It has been a couple of years since I built and delivered an amplifier to my brother. I thought it was a pretty novel, simple and effective circuit. I was in a rush to give it to him as a Christmas present, so I didn't get all of the time that I wanted to explore and optimize the topology.

Well, it seems that it broke and so I took it back to repair it. The problems were determined to be inadequate heat-sinking on some SMT power resistors. I fixed the problem and decided to spend some time testing, optimizing, and upgrading. For reference, here is the schematic of what I initially delivered:

The first measurement I made was output impedance at a few different output tube idle currents (450V B+). Here are the results:

3.7 Ohms @ 50mA
3.3 Ohms @ 60mA
3.1 Ohms @ 70mA

Those are a bit higher than I liked. I also noticed that the 10k SMT resistor in the output tube feedback network was probably running a little hot for the heat-sinking that the board offered. I had also experimented with similar feedback on EL34s and had found that a 100k feedback resistor should be more than adequate in this position, saving some heat, so I did some analysis and decided to decrease the current in the feedback network and increase the feedback from 20% to 30%. The -388V rail for the p-channel fet should have no problem supporting that and the 6BL7 should have adequate voltage swing as well. The amp also has more gain than desired so this increased local feedback should be good all around. Hopefully, this would get output impedance down somewhere close to 2 Ohms as well. The Hammond 1650R output transformers have a primary DC resistance of 97.8 Ohms and a secondary resistance of 0.21 Ohms, so amplifier Zout is still dominated by effective output tube rp and not high DCRs in the output transformer. 

Other changes include modifying the input to accept balanced signals and adding a bias servo to keep output tube idle currents precisely controlled at all times. Here is the modified schematic:

After changing the feedback resistors, the Zout were measured as:

2.9 Ohms @ 50mA
2.4 Ohms @ 60mA
2.2 Ohms @ 70mA

I took detailed distortion measurements at many power levels at all three idle currents and was surprised at how little variation there was, so I decided to bias the output tubes at 60mA. It seemed to be a good balance of lowering Zout but getting long output tube life. Here are some of the distortion spectra:










20Hz square wave

1kHz square wave

10kHz square wave

The finished product.

Overall, I think this is a great result on this amp. I think it's a pretty sound approach.

Friday, January 27, 2017

A Test Amplifier for Tube Circuits

Have you ever just needed to test a driver-type tube stage but your sound card output can't put out enough voltage drive to get it where it needs to go? Well I have, and here is how I solved my problems.

I searched around and looked at high voltage op amps and found a few candidates. The one with the lowest distortion and noise specs that I could find was the ADA4700. I built a simple non-inverting stage with a gain of 20 and a 100k feedback resistor, powered by a couple of medical isolated 48V wall-wart power supplies (Meanwell GSM25U48-P1J) and got some pretty good performance. See below:

The above results are for 10Vrms, 20Vrms, and 30Vrms respectively. I was pretty happy with that performance, but I was getting more distortion than I expected from reading the ADA4700 datasheet.

Later, I was reading a small signal audio design book by Douglas Self and he brought up the point that some (most?) op amps suffer from much higher distortion when used in a series-feedback configuration than if used in a shunt-feedback configuration. He notes that the LM4562 is a notable exception.

I wanted to have the high input impedance of a series feedback configuration but the low distortion of shunt-feedback so I decided to try a two-stage version with an LM4562 inverting amplifier as the input stage and the ADA4700 inverting amp to get the voltage output up to 30Vrms. Particularly, I wanted to minimize higher-order harmonics so that I could recognize them in the tube stages that I wanted to test.

Here is the schematic of what I came up with:

I had the idea that I could put filter capacitors across the feedback resistors and that might have some positive effect on suppressing higher order harmonics. I added a bit of gain to make up for the loss and set the -3dB point at 1kHz for the filtering of each stage. Here is a picture of the circuit as built:

The ADA4700 only comes in a surface-mount package with a thermal pad on the bottom so I bought a prototype adapter, bent the leads, and flipped it on its back. I make a makeshift heatsink out of a couple of pieces of solid-core wire. I don't think this is really needed at the loads I am driving but it was easy to do, so I did it. I stuck with a surface-mount LM4562 because it fit well right next to the ADA4700 on the prototype board. The filter capacitors are placed in parallel to the feedback resistors and easily soldered/unsoldered on one side to take them in or out of the circuit. How does it perform?

First, here is what the distortion of my test setup looks like (including an M-Audio M-trak II plus and a high-impedance input version of Pete Millett's soundcard interface):

Here is distortion of the LM4562/ADA4700 circuit at 10Vrms, 20Vrms, and 30Vrms output, without the filter capacitors in circuit:

I was happy with the reduction in distortion of the new circuit. Let's see what happens with the filter capacitors in:

The first thing I noticed is that the noise floor is significantly lower at frequencies above mid-band. This could be helpful in detecting small levels of higher order harmonics in testing. As you can see, THD is also slightly lower. The main reduction comes from suppression of 2nd harmonic. Third and higher harmonics seem to come up a little bit. Noise floor seems unaffected in the 30Vrms plot but that is because I have to switch ranges on my Pete Millett sound card interface to a setting with a higher noise floor. The sound card interface becomes the dominant noise source in that plot.

I think I will stick with the unfiltered configuration mostly but I can see situations arising where I might want to put the capacitors back into the circuit.

Saturday, May 9, 2015

A Push-Pull Amplifier with Simple Plate-Grid Voltage Feedback Driven by a P-Channel FET

For this post, I want to go back to some stuff I did in December 2013. Some time before that I read O. H. Schade's paper Beam Power Tubes and was fascinated by the characteristics shown in that paper of what plate-to-grid voltage feedback was able to do to the characteristics of a 6L6. In fact, those characteristics looked more linear than a 300B. A 6L6 could never replace a 300B, because it doesn't have a high enough plate dissipation rating, but the linearity was impressive nonetheless.

I started to wonder what KT88 (which does have a comparable plate dissipation rating) characteristics would look like with plate-to-grid voltage feedback. I followed the procedure in Schade's paper to plot curves in Excel, which are shown below:

I decided that in order to do a fair comparison between a 300B and the KT88 with feedback, I needed to plot on the same area. In the plot above, the pinkish purple plots are KT88 with feedback with the feedback resistors set for a mu of 5. The Yellow curves are 300B which has a mu of about 3.8. The Cyan curves are KT88 triode connected which has a mu of about 8.

As you can see, the 300B is more linear and has a lower rp than a triode-connected KT88 but has a lower mu. But the thing that is really neat is that the KT88 with plate-grid feedback has higher mu, lower rp, and better linearity than the 300B. We have the KT88's high transconductance to thank for that. It also has the benefit of being able to reach saturation without driving the grid positive, so we can drive the plate to a much lower voltage without needing a driver to supply grid current. All of the pinkish purple curves represent negative grid voltages. Needless to say, this got the wheels spinning in my head. I was considering building a 300B amp some day but they are kind of expensive tubes. KT88s are cheap by comparison.

Now I had to figure out the best way to apply the feedback. Most of the practical designs involved capacitive coupling of some sort. I really like to keep anything that goes to the grid of a power tube direct-coupled if possible so that if the amp is over-driven, there will be no cap to charge and cause blocking distortion. There are ways to minimize this but I saw an idea in a post one day on the tube amplifier forum at that showed a totally simple idea of how to apply plate-grid feedback, using a p-channel FET just like this:

It was simple and elegant. I was in the process of building a push-pull KT88 amplifier for my brother so I altered the design a bit and came up with this:

Pictures of the construction/testing:

I was desperately trying to finish this amp in time to deliver it to my brother as a Christmas present so one of the things I regret is that I didn't take many measurements. I remember it put out almost 50W at clipping but I only saved screen captures of 1W and 10W distortion measurements. This was also the first time using my measurement setup based on Pete Millett's buffer box. I built a high-impedance input version so that I could take measurements with a X10 or X100 scope probe directly on tube plates. It works well but is much more susceptible to noise pickup. I later spent some time calming the noise down in the measurement setup, so please excuse the noise peaks below the fundamental. I also later discovered that my diodes that I was using to protect my sound card input were generating high-order products at higher amplitudes, so some of the high order products in the second plot may not in fact exist.

Here is distortion at 1W:

Distortion at 10W:

Curiously, even though I put that AC balance pot in the input stage to null out gain mismatches between the two sides of the 6BL7, I never was able to get the second harmonic on the output of the amplifier to be lower than the third harmonic. I wasn't expecting that.

Anyway, it sounds awesome, my brother couldn't be happier, and I don't think I'll ever build a 300B amp.