The "absolute" cancellation of even ordered harmonics never happens in a real P-P amp. For perfect cancellation the drive levels, OPT halves, and output tubes must be perfectly matched over all operating conditions. This doesn't occur in the real world.
Knowing this, is it possible to purposely create imbalance to alter the cancellation level. I have an old Scott 272 stereo amp that has an AC balance control in the PI. Guess what this knob does? It alters the harmonic signature of the amplifier by allowing perfectly matched drive levels between the two output tubes, or a less than perfect match. The manual suggests adjusting this pot for lowest distortion, but I found that I liked some intentional imbalance.
These experiments over 10 years ago led me into a 3 or 4 year long case of SE fever, from which I didn't really recover until I built a P-P 300B amp.
The SE amps have a detail and imaging that just doesn't seem to be as good on a P-P amp. P-P amps have dynamics and transient capability that isn't there on most SE amps, and bigger more power full SE amps are not the answer. Is there an amp that does both perfectly.....I don't know, I haven't found it yet. I have built some pretty nice sounding P-P amps, and all involve some intentional imbalance in both the DC and AC settings.
Note that sometimes enough imbalance is achieved with mismatched output tubes, and a slightly different bias setting between the 2 tubes. This can get the 2H up to, and even above the 3H.
But the listeners can set it up so that they get the best of both worlds. Active-biamp (SE for high & mid, P-P for low).
I did some experiments along these lines. My limited experiments say no. I have explored this avenue mostly from the viewpoint of obtaining higher output power than I could squeeze through a single OPT of a given power rating. The results may be different with lower power.
I tried multiple ways of simply connecting two OPT's together to a common amp and this proved to be a compromise...always.
I then experimented with multiple amps, paralleled or bridged, like I have stated in my earlier posts, with some good success.
My latest, and best efforts have centered around a common input and driver stage feeding multiple output stages, each with its own bias and OPT, and then combining the outputs of the OPT's. This even works with local or global negative feedback.
Which PP amp sounds better than which OTL?
My DIY 6CA7G PP amp was kicked ten fold by an Atma-Sphere S30 OTL. The OTL have SET magic and PP muscles at least in my system. So it's a keeper.
Oh, and the OTLs from Atma-Sphere has for each audio channel, two single-ended triode amps running 180 degrees out-of-phase with each other (differentially). The output is formed around a circlotron coupling.
I have designed and built SETs, OTLs (160 watts RMS, triode configuration), other OTLs, and designed my PPs around those, ABing of course.
Tubelab, I already figured the phase splitter was the best way to go. Thanks for your input.
Cheers.
Sorry, but such description isn't helpful. "better" is subjective to the listener's preference. It may only work if someone is writing to his / her identical twin brother / sister.
"Better" is also highly dependent on the type of music played, the mood of the listener, and the desired volume level at the time, even with the same listener.
I have been building audio amps since I was in high school (over 40 years ago) that used tubes, silicon, and chips. I have probably built at least 100 over the years, and I still have a few.
If I want to chill out to Norah Jones or something similar, I will hook up my 45 based TSE or a KT88 based SSE because they just sound right.
If I want to crank some Pink Floyd or Depeche Mode, then I will hook up a P-P amp. Which one depends on how many neighbors I want to annoy.
If I want to recreate a Metallica concert in my living room (except for the pyro) it has th be the 125 WPC sweep tube based P-P through 96 db 15 inch coaxials!
Right now the Lollapalooza concert is streeming live through a 32 inch TV and a Simple P-P courtesy of Youtube/Google
I read of an amp builder running 2 SE output transformers on a PP amp with the primaries in series which he really liked if he spent time finding closely matched transformers. The gap on a SE trans didn't matter when run this way.
As for SE or PP maybe class A PP with DHT tubes is a good compromise?
As for SE or PP maybe class A PP with DHT tubes is a good compromise?
Last edited:
Best OVERALL sounding amp I ever made was a 300B P-P. Not quite as detailed as the 45 SE, and not as powerful or as punchy as a big powered sweep tube P-P, but did a good job at either end of the music spectrum. Too bad it blew up.
Now that I thought about it, I have several Transcendar 300B SE OPT's, I'll try wiring 4 into a Simple P-P amp.
Better in what way, measurement, audible transparency, price per power output, aesthetics?
That's the important aspect. If you like what you hear, that's great. But that may not be the same to others. Different people have different taste.
And others like PP over SEs and OTLs as well.
The problem with DHT tubes is two fold.
The cathode circuit is usually overly complex unless one goes AC. If one thinks it is hard to find a good sounding part, then the problem is multiplied in the cathode circuit unless AC is used.
Secondly, the drive signal requirement is ridiculous for 300bs etc. And 45s hardly put out any power. Combine the higher distortion from the driver tube developing the large drive signal and one creates more higher orders of distortion.
All and all, over the decades, I have never heard a dht sound right.
Cheers.
The problem with DHT tubes is two fold.
The cathode circuit is usually overly complex unless one goes AC. If one thinks it is hard to find a good sounding part, then the problem is multiplied in the cathode circuit unless AC is used.
Secondly, the drive signal requirement is ridiculous for 300bs etc. And 45s hardly put out any power. Combine the higher distortion from the driver tube developing the large drive signal and one creates more higher orders of distortion.
All and all, over the decades, I have never heard a dht sound right.
Cheers.
Nah...tie one pin of the heater directly to ground....feed the other with a regulator chip. CCS or constant voltage....your choice. AC heating will cause measurable IMD of the signal with line frequencyand its harmonics. It is usually too low a level to hear directly, but causes listener fatigue and a vaguely dull sound.
True, the popular low Mu DHT's take lots of drive voltage, and they will start to draw grid current before the grid actually goes positive. Nevertheless it is relatively easy to generate the needed drive. Many triodes like the 417/5842 and the good old 6SN7 can generate plenty of drive VOLTAGE with near zero distortion if fed properly. Load the triode with a CCS connected to a high voltage source, and buffer its output with a follower so it is working into a near infinite load. A tube purist can build a CCS and a follower with tubes, I prefer silicon because it just works better. I have an 845 SE amp that makes 40 WPC in A2. It is driven by a 45 loaded with a CCS from a 500 volt supply and buffered by a mosfet follower. It can swing a 300 volt P-P signal into the 845 grid with a very low distortion.
A 45 in A2 SE will make about 2 WPC. This is enough to feed many high efficiency speakers, but I run it into my 87db Yamahas and it is still loud enough for casual listening with most music.
A pair of 45's in AB2 P-P can make 18 watts according to the old RCA data sheet. I breadboarded such an amp and saw 20 WPC but I was running the 45's over the plate voltage spec. The breadboard revealed a very nice sounding amp that will get built some day.
I prefer my 6CA7 PP over the 300B SETs I heard in my system even an Audio Note 300B PSE sounded colored, undynamic, lifeless and lacking both ends of the frequency spectrum.
But the OTL is so much better than my PP amp so there's no way back.
Sounds good on the surface, but first for newbies, the positive portion of the filament/cathode is where max cathode current is concentrated since the grid to cathode voltage is least negative. As such, the positive side of filament should be grounded, regulate the negative side to minimize power supply/regulator sonic signature. Even with negative portion regulated, regulator parts, chip is in the circuit path and contribute to sonic artifacts being created.
How much distortion? Near zero at 120+ volts peak to peak? Wishful thinking.
Why drive the grid positive when one does not have too. Again superficial as the SS parts will impart a sonic signature. Tube has higher output Z, thus larger kink. Adding another stage imparts more frequency dependent feedback from stage to stage through the power supply.
Casual listening is about it. Again a directly heated cathode with its associated parts and problems. And high efficiency speakers have problems reproducing the entire audio band distortion free. Maybe external powered sub as well?
Glad you like the sound, but maybe not necessarily what others like. Still the problem of parts in the cathode circuit. If an electrolytic capacitor is a problem, other parts are not good either.
Cheers.
Last edited:
OK, 0.3% is near zero in my book. Yes, the tube must see a near infinite load, have a high headroom voltage, and you need to try a handfull of tubes.
I see about 0.7% at the grid of the 845 at 250 volts P-P. It is driven by a 5824 followed by a 45 each CCS loaded and mosfet buffered. I have spent some time perfecting big voltage drivers for screen drive and cathode follower output stages.
Agreed, some tubes like the 45 have 50 volts or so of bias and a 2.5 volt gradient across the filament, so this is not a big deal. The 811A however is a different story one end of the filament is conducting all the current while the other end is cut off.
When I first starting making the Tubelab SE I heard similar comments some even suggesting I change my name to transistorlab. After a few years and a few hundred amps, I don't hear any complaints about the sound quality, or sandy artifacts. The right part in the right place in a well designed circuit makes the difference. Yes, I understand the VVC effects and their ability to cause PIM.
If all the driver stages are connected to the power supply through its own CCS, and each CCS has at least 1 megohm of dynamic impedance from DC to 100KHz, there is no coupling through the power supply.
This thread was started to explore the possibility of making a pseudo push pull amp by connecting two SE's together. It has wandered into the realm of which amp is "best" which has been debated forever, and won't be answered here. Its safe to assume that there is no one "best" amp, or we would all have one already. I, as a single user can't even pick out my own "best" amp since I like the individual merrits of each type. Yes, I have played with OTL as well, but I haven't made one of them that I like yet. Heat loss is a big deal here where we need to run the air conditioner in winter. Thats why the 845SE hasn't been turned on in 4 years.
I'll report back when I have had the time to stuff some SE transformers into a P-P amp, or wire two SE channels together.
I, personally, would not call 0.3% (~ -50db) or 0.7% (~ -43db) near zero. My drivers typically have .05% (~ -66db) distortion at 200 volts P-P, so no chance of higher order products getting in the way. The higher the order, the higher the weighting factor vs 2nd harmonic.
In either case, the SS artifacts can show through, depending upon concealment from masking distortion.
I am glad no one has commented on yours. However, what are they referencing to?
A couple of problems with that reasoning. First, SS have bleedthru problems. Even a 1n4148, reversed biased, with 1pf, has problems. Secondly, all the added parts create their own sonic problems. Thirdly, one can still have feedback problems to the output stage.
I have to disagree in all points.
There is no "debated forever". OTLs have high output impedance (Z) and thus needs a high Z speaker. Even then it has problems with deep bass response.
SE amps have the same problem (not the same solution), although parafeed is better. High frequency response problems also seem to be extent. However, choice of tubes helps in this regard.
Combining two SEs has problems in matching as well, and it is inconsistent from unit to unit because of transformer and other part tolerances.
If one checks over properly designed PP amps, all the problems can be improved upon, and consistently. Only bass may be lacking and that is vs SS designs. (Although SS designs can be over bassed.)
Cheers.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.