Premium Sweetie Amp: Technical Talk, Parts Options, Testing Results and Other Comments

[henry4841]

I decided to check how much was distortion in the .66% and .67% eliminating the noise number and the figures are .56% and .54%  for just distortion, which is dominant 2nd harmonic by the way. The kind of distortion one expects to see in a SET amplifier. I would expect some chokes in the PS section would make it look much better on this Arta software but why would you. This amplifier is quite on my LaScala's with my ear next to the speakers and it sounds wonderful to my ears just as it is. Also curious on what Maynard thinks is the cause of the peak between 300hz and 700hz. Looks to me like that is what is carrying on further up the frequency range. 

 

Fascinating not seeing any 3rd harmonics to speak of on this amplifier on Arta. Curious to hear what Maynard thinks is why. This maybe is what makes this amplifier stand out from the crowd of SET's.

[henry4841]

Looking at the screen shot of the Arta on the Sweetie I honestly believe a choke in the PS section would clean up the hump at 500hz. I looked to see if I had any chokes in my stash and I do not. Used my last one on the KT-88 build. The cheapest ones I see at I could use at Mouser are $13.40 apiece raising the cost $26.80 to make this amp look better on the bench and during test with no benefit to my ears when listening to music. For curiosities sake I may purchase a couple of chokes just to satisfy my curiosity. I do not think there is a problem in the build or parts being this amp is built as mono's and both channels look almost identical on the screen. Look forward to hearing Maynard's opinion.  

[captainbeefheart]

That's quite a lot of 60Hz being modulated into the 1kHz fundamental.

 

60Hz is literally only -10db below the second harmonic. I'd say before looking into chokes figure out where all that 60Hz is coming from. Once you clean that up then we can see how much of the power supply ripple is actually getting through.

 

As for distortion with the Sweetie, I'll have to double check through my folders because I thought THD was much higher at 1/4watt output. I only mocked up one channel to my variable DC bench supply.

 

I just want to make sure you are measuring the correct output amplitude into 8 ohms for 1/4 watt

 

1/4 watt = 2v peak output into 8 ohms correct?  I just wanted to make sure we are on the same page in regard to peak voltage across the load while looking at ARTA.

 

1/2 watt = 2.8v peak

 

1 watt = 4v peak

[henry4841]

 

I made a mistake calculating 1/4 watt into 8 ohms into voltage at resistor. The correct voltage for 1/4 watt into an 8 ohm load is 1.414v's and I was using a much lower voltage previously. Sorry Maynard. My only excuse is never tested using Arta at 1/4 watt. That aside with the correct 1/4watt into 8 ohms the graph looks like this. For a 1 watt amplifier this graph looks great to me. One has to remember the 6Y6 tube is a very old tube design when 1 watt was the norm. As I have stated many times about the Sweetie it has a tone that sends me back to the era of my childhood. The sound from a backelite radio my mother played while doing her morning housework. Lovely sounding little amplifier with just enough power for any of the Heritage speakers. For most that is. 

 

 

[captainbeefheart]

Thank you for double checking your measurements. That looks pretty much spot on to what I was measuring - 1.42% THD at 1/4watt.

 

With the larger fundamental frequency now the 60Hz noise looks much better.

 

 

 

 

[captainbeefheart]

The amp has plenty of gain for just 1 watt of power; ~17db

 

I added a 2.2k resistor from the Sweetie's speaker output to the 6SJ7 cathode which gave about -7db of feedback. It cleaned it up a lot but not sterile clean by any means. I didn't write it down but remember around 1 watt at 1% THD.

 

So with one resistor you can get less distortion at 1 watt than you are getting at 1/4 watt output now. An input of 1v will bring you to 1 watt output.

 

*Note: I was using 560 ohm cathode resistor for the 6SJ7 and it wasn't bypassed with a capacitor.

[henry4841]

The error I made is in my math finding 250mw at 8 ohms. As most on this forum know 2.83vs into 8 ohms =1 watt. The number is what you see used to measure efficiency of a speaker at 1 meter.  So I just used simple math thinking if I divide 2.83vs by 4 I will have the volts needed at my 8 ohm load to achieve 1/4 watt, wrong. That is the amount of voltage, .7 I was raising the volume control to see at 8 ohms thinking I had 1/4watt. I should have used ohms law and I would have found that 1.4vs at 8 ohms was what was needed to achieve 1/4 watt, 250mw. In other words the old previous graphs were at .06 watts instead of .250 watts. Just a simple ohms law calculation error on my part.   https://ohmslawcalculator.com/ohms-law-calculator  

 

I had rather not use feedback in a SET amplifier. Sure it will produce better numbers but at what cost. I have nothing against feedback, just not in a SET.

[captainbeefheart]

9 hours ago, henry4841 said:

Sure it will produce better numbers but at what cost.

 

At what cost? 

 

The cost is -7db in gain. Which isn't enough to cause stability trouble if that's what you are worried about.

 

9 hours ago, henry4841 said:

I have nothing against feedback, just not in a SET.

 

I find this fascinating. What's the difference? Why can't one use feedback with a SET amplifier?

 

 

[henry4841]

2 hours ago, captainbeefheart said:

 

At what cost? 

 

The cost is -7db in gain. Which isn't enough to cause stability trouble if that's what you are worried about.

 

 

I find this fascinating. What's the difference? Why can't one use feedback with a SET amplifier?

 

 

Sure you can but it is not something I care to do. Just me being old school probably. 

[henry4841]

54 minutes ago, henry4841 said:

Sure you can but it is not something I care to do. Just me being old school probably. 

That and the only significant distortion I see on the graph is 2nd harmonic so with feedback you will be reducing the amount of 2nd harmonics. I like the sound of this amplifier probably because of it's 2nd harmonic signature. Also I think the lack of the typical touch of 3rd harmonic that is absent in this SET contributes to it's pleasing sound I like as well. All in all at 250mw I like what I am seeing on the graph. 

 

Certainly if one wants to play some with this 1 watt amplifier you could install a switch and change back and forth between feedback or not. You would have to play around with the amount of feedback till you are satisfied with what you see and hear. 

[captainbeefheart]

1 hour ago, henry4841 said:

That and the only significant distortion I see on the graph is 2nd harmonic so with feedback you will be reducing the amount of 2nd harmonics. I like the sound of this amplifier probably because of it's 2nd harmonic signature. Also I think the lack of the typical touch of 3rd harmonic that is absent in this SET contributes to it's pleasing sound I like as well. All in all at 250mw I like what I am seeing on the graph. 

 

Certainly if one wants to play some with this 1 watt amplifier you could install a switch and change back and forth between feedback or not. You would have to play around with the amount of feedback till you are satisfied with what you see and hear. 

 

I'm  very curious what type of music you listen to?

 

I'm not in the camp of thinking that distortion must be as low as possible, but I do believe there is a point of too much even if it's only second harmonic. With simple styles of music very high second harmonic distortion isn't as much of a problem, it's when you get complex signals and demanding program material where too much distortion just doesn't work due to the amount of intermodulation distortion. Basically the harmonics don't play nice with each other with certain types of music.

 

There are merits well beyond the distortion aspect of negative feedback, the control of the speaker with the lower output impedance is a big advantage along with reduced noise floor are all welcome in my book.

 

I don't think one needs to play with the amount of negative feedback with the Sweetie because there just isn't enough gain to do so. Using more than -7db and you will require more than your average source can deliver to reach full output power and anything less than -6db of negative feedback isn't really going to be much worth it as it will just pretty much sound the same as with no feedback.

 

That's the best part about Triodes, they don't require gobs of negative feedback to reach excellent numbers, but I do think they need some in most cases. If someone has say K-horns or La Scalas and they listen to quite simple music at moderately low levels than a simple triode amp with no feedback can get you by or even sound great with all that coloration. But once you start getting a full orchestra or some slamming Electronic Dance music especially with direct radiating speakers like a Forte, Heresy, or Cornwall then the amp needs control (low output impedance)  and low enough distortion to sound good. Although my La Scalas and K-horns seem to do better with these high output impedance amplifiers they still perform better with a more controlling amplifier.

 

There are two aspects where these types of amps produce a mirage of "good bass". One is the resonant frequency down low that gets agitated and so the listener is tricked into thinking this "one note bass" is superior. Then there is the fact that the output transformer starts to saturate at low frequencies giving rise to even more excessive distortion. Since this type of distortion is very non-linear it gives rise to even harmonics. The second harmonic of say 35Hz is 70Hz, this new signal of 70Hz that wasn't there before reinforces the 35Hz fundamental frequency. I have seen over 30% THD at 35Hz with some of these amps and it's amazing that it still doesn't sound terrible. 30% THD is not hifi no matter how well masked it is and although it will sound novel and exciting at first, for me it gets old quickly. I have seen people swap out these type of amps for SS amps and say the SET amp had "better" bass. But once their ears gets used to accurate reproduction with tight bass it becomes apparent the average SET amp in many cases is a 1 trick pony that only does simple music well.

 

Just my opinion so please don't take any offense. I find that people that prefer these "no feedback" amps listen to very simple music at lower volumes.

[henry4841]

13 hours ago, captainbeefheart said:

 

Just my opinion so please don't take any offense. I find that people that prefer these "no feedback" amps listen to very simple music at lower volumes.

None taken. Doing what sounds best to you is the best part of building amplifiers. As you may know I have lots of amplifiers, around 25 working ones both SS and tube. Most are class A, my preference. My music choice does play a role in what I enjoy but most that come to my house and listen with me consider my listening level is too loud. I am definitely in the league of simple circuits with few parts and no feedback. With no feedback in a SET amplifier you are listening to what the tube sounds like without any manipulation. I have plenty of SS amplifiers that do have feedback in their design and for a change I can enjoy them but my preference is a good SET amplifier without any feedback. For rockers and modern music one would probably prefer some feedback but then they are not a large market for a 5 to 7 watt SET amplifier. Most of them look for mega watt amplifiers with horn speakers needing only a few milliwatts of power to play too loud. Right now class D is all the range and I only expect it to be the norm for the masses in the future.  

[captainbeefheart]

5 hours ago, henry4841 said:

None taken. Doing what sounds best to you is the best part of building amplifiers. As you may know I have lots of amplifiers, around 25 working ones both SS and tube. Most are class A, my preference. My music choice does play a role in what I enjoy but most that come to my house and listen with me consider my listening level is too loud. I am definitely in the league of simple circuits with few parts and no feedback. With no feedback in a SET amplifier you are listening to what the tube sounds like without any manipulation. I have plenty of SS amplifiers that do have feedback in their design and for a change I can enjoy them but my preference is a good SET amplifier without any feedback. For rockers and modern music one would probably prefer some feedback but then they are not a large market for a 5 to 7 watt SET amplifier. Most of them look for mega watt amplifiers with horn speakers needing only a few milliwatts of power to play too loud. Right now class D is all the range and I only expect it to be the norm for the masses in the future.  

 

How do you know you are really hearing the 6Y6 in a no feedback circuit? The 6SJ7 is supplying the input signal to the 6Y6; I ask is the signal taken from the 6SJ7 plate a perfect copy of the input signal? It's not, with the Sweetie producing 1 watt output the 6SJ7 already is producing 1% THD from it's output alone. So you are never hearing the 6Y6, you hear the whole circuit which in these amps the driver can produce as much distortion as the power tube.

 

Since the global feedback doesn't just clean up the power tube distortion but also the driver, you may actually be hearing the 6Y6 more with feedback since the 6SJ7 will produce a much cleaner output to drive it. E.g. The signal isn't already tainted with 1% THD before even reaching the 6Y6. So how can you give the 6Y6 even a fair listen when it's input signal is already heavily distorted?

 

Driver stages in two stage SET amps are extremely important, almost more so than the output tube. I spent a lot of time on driver stages for SET amps and high gm pentodes strapped as triodes work very well when CCS loaded.

 

 

I am one of the few people I suppose in the market for a SET amplifier that can do all music genres well. The great part about triodes is they require very little feedback to get fantastic results. Try and remember feedback is not bad, it's why triodes behave the way they do. It's when designers don't properly apply feedback or just use it as the only means of linearity. It's not magic, it can't make everything good, but when you have a great start with triodes just a tad will make a good amp a great amp.

[tube fanatic]

Captain, did you measure any phase shift in your breadboarded Sweetie with fb?  Some listeners seem to be able to hear it with certain music so I side with Henry in not using any with SETs.  As far as simple vs. complex music goes, I like to use large choral works to make some evaluations.  In years past I compared the tube amps with SS units having virtually no distortion (but large amounts of nfb of course) and heard little or no difference on such selections.  Congested sounding vocals were still that way with the SS amps.
 

At this point my conclusion is that it really doesn’t matter how the amp measures if the listener doesn’t like the sound.  So called “accurate” reproduction is not always the most pleasing to listen to.  There’s a reason why so many use the tone controls on vintage amps and preamps so liberally.

 

Maynard

[captainbeefheart]

If you really want to hear the 6Y6 in triode mode as clear as possible with no feedback I highly suggest swapping out the 6SJ7 for a cleaner driver circuit. Don't get me wrong I use the 6SJ7 and like it a lot but in the Sweetie it's gain is a tad over 50 which is fairly low for this tube.

 

Triode strap something like a 6EJ7 with a 15mA CCS on the plate and Rk = 220. You'll get more gain and at 60v peak to peak output it's only around .2% THD.  Now that's a very clean driver for a no feedback amplifier.

[henry4841]

26 minutes ago, tube fanatic said:

At this point my conclusion is that it really doesn’t matter how the amp measures if the listener doesn’t like the sound.  So called “accurate” reproduction is not always the most pleasing to listen to.  There’s a reason why so many use the tone controls on vintage amps and preamps so liberally.

 

Maynard

It's entertainment not rocket science. I am sure that not everyone would take to the Sweetie with it's generous amounts of 2nd harmonics. For me it is what gives it it's charm and makes it different. Nelson Pass when discussing First Watt products often says he just wants to hear the transistor clapping. Much could be said for those that like simple SET tube amplifiers with few parts. I love the sound of that old 6Y6 tube. By the way, it is surprising to me how good the bass is for such an anemic little amplifier, 1 watt. 

[captainbeefheart]

4 hours ago, tube fanatic said:

Captain, did you measure any phase shift in your breadboarded Sweetie with fb?  Some listeners seem to be able to hear it with certain music so I side with Henry in not using any with SETs.

 

I fail to understand the question.

 

In amplifiers the feedback fraction B is constant; it does not change in either magnitude or phase as we change frequency. The internal amplifier gain A is the part that changes with frequency and produces the phase shift.

 

All amplifiers possess reactance, the coupling capacitors produce phase shift at low frequencies while stray capacitance in plate circuits cause them in high frequencies.

 

Phase shift is already present in your Sweetie Amplifier along with every other open loop amplifier. 

 

4 hours ago, tube fanatic said:

Some listeners seem to be able to hear it with certain music so I side with Henry in not using any with SETs.

 

If phase shift is an audible problem then that would actually be an argument FOR the use of negative feedback since it reduces phase shift in the amplifier.

 

In order to not have phase shift in the audible frequency range one would require the amplifier to have very wide bandwidth;  negative feedback increases bandwidth and so reduces phase shift in the amplifier.

 

 

[captainbeefheart]

Negative feedback reduces phase shift

 

 

[captainbeefheart]

Since the Sweetie designer is asking me about his circuit with feedback I feel I still owe an answer to his original question.

 

I already gave some distortion figures of the Sweetie with and without the feedback.

 

As for why I added the feedback originally wasn't only for the distortion aspect, it was also because I was getting a roll off well before 20kHz. I assumed it was the cheapo universal output transformer I was using wired for 5k:8 operation.

 

Without any feedback phase shift at 20kHz was already -80° and with the 2.2k feedback resistor added it reduced phase shift at 20kHz down to -40°.

 

-40° phase shift at 20kHz isn't great but it's a lot better than -80°. Compared against another amplifier I made with some decent output transformers and -20db of feedback I was getting only -20° phase shift at 20kHz. This was a 3 stage amplifier also which of course is going to have greater phase shift from the extra stage.

 

 

I think there are two current builds I have seen of the Sweetie, one with Hammond OPT and the other with Edcor. I would be very curious to see what the phase shift in your builds are. A] Because of the better OPT; and B] Probably better layout compared to my breadboard. I had flying leads and so my stray capacitance was most likely worse than a tidily built amplifier layout.

 

 

 

 

[tube fanatic]

Sorry Cap, but your arguments are only partially correct.  But, this is not the venue for such a technical discussion.  If I can find time during the week between Christmas and New Year, I will start a thread on DIY Audio and discuss it with you there, where it belongs (I assume you are still very active on there as well).  We can post a representative amplifier and analyze every factor from input to output and look at what happens everywhere.  Then, we can have the input of other EEs to really spice things up a bit!  

 

Maynard