The Singing Bush Tips 'n' Tricks

Thanks.

Funny thing is, in this case, the cables measured fine. The lesson I learned is to not doubt the obvious problem is the problem even when measurements suggest otherwise.

If only I had made another cable months ago - all those months of silence lost ...

Take care,

PS I cannot help but wonder if the bias supply could not be made better with a fancier regulator? Whether it would make a sonic improvement is something that haunts me. The gist of the audio sickness - even when one is hearing something better than they have ever heard before you still want another dose of that great rare moment when an improvement is an improvement, a real improvement. As dangerous to one's wallet as heroin.
 

Ben Mah

Member
Paid Member
2012-05-18 6:36 pm
Vancouver
I think adjusting the operating point which affects the distribution and amount of harmonic distortion would have a much bigger affect on the sound than changing the bias supply regulator.

Fun can be had by measuring the baseline harmonic distortion and then adjusting the operating point to change the distortion profile, and then listen and decide whether it is an improvement.
 
Your advice is taken very seriously.

I assume the distortion distribution would likely be system dependent. With four amplifiers this could be a serious endeavour!

When you say the operating point is this simply moving the Iq and Vgs about? One can arrive at the recommended values with many different trimpot settings. I have wondered if there is one best or are they all about the same once you get to a certain chosen pair of values?

Would this be changing the recommended values or finding the best spot for each trimpot for the recommended values?

I hope that makes some sense.

I wonder about the phase thing concerning the distortion profile. I can easily change the polarity in my system with the input signal but is this business about the polarity based upon how the output is fed to the speakers?. To find the best position would one change the polarity of the speaker wires and then make the appropriate change to the input signal to bring it back or would simple switching the polarity of the input result in the benefits described?
 

Ben Mah

Member
Paid Member
2012-05-18 6:36 pm
Vancouver
Yes, I mean Iq and Vds. I think 2.5A is good for THF-51S. Changing Vds probably has the biggest effect so I would play with that. With my mu follower amps I also play with the location of the speaker output at the resistors between the mosfet and the SIT. However the Singing Bush board does not allow that to be easily done.

Trimpot values are not that important as long as the trimpot adjustment changes the operating point to where you want it. Unless the Tokins are matched, the trimpot values will be different for each Tokin.

In order to get a sense of what is happening, the harmonic distortion should be measured in order to get a sense of what happens when the operating point is changed.

As for distortion phase, I switch speaker wires around and choose what I like best. I don't change the amplifier wiring. That way I know that the speaker negative is ground so if I use an oscilloscope I don't mess up.
 
  • Like
Reactions: 1 user
As usual being late to the party, I got some Tokin SITs (THF-51S) before all SITS disappeared from reliable sources. Impatiently waiting for Pa to reveal circuit for these Toking Biguns, I started trying to figure out what to do with them. Scouring diyA I considered all sorts of options (mostly from ZM and Ben Mah).

Follower seemed boring and a waste of this device, Active loading will be hot as hell (especially summer time in Texas). Hmm... leaning towards choke load.

Lately my favorite amp to listen to has been a Choke Loaded Lu'amp (basically Pa's Zen v8 with choke replacing the lightbulbs for load and half the power supply voltage). There is something just really likeable about this amp's organicity.

:idea: I'll just do a SIT version. This SIT Amp will end up being even simpler than that one.... so I thought.

More reading of threads later I realized I need a buffered bias voltage... Aha :scratch2:
I could use the lower half of Singing Bush and done... Then I read this:

Tokin SIT Warnings.png


Ok so I need power sequencing. ON Bias Supply -> Drain Supply..... OFF Drain Supply -> Bias Supply
This is getting even more complex. I even considered a Power Switch + Standby Switch to keep things simple (like old solid state rectifier tube amps have so plate voltage does not hit power tubes before they warm up).

Mighty ZM to the rescue! Choked Singing Bush solves all problems. I was trying to reinvent the already ZM invented wheel (no wonder I'm a ZM fan :worship:).

I should have read the whole thread.

The big question for you, Zen Mod, is when will these boards become available?
No hurry, just wondering when this iron comes out of the Zen Mod fire... there are so many irons in that fire all at once. :rofl:

Cinco
 
  • Like
Reactions: 1 user
Ben Mah - I think polarity makes a difference in the sound for different recordings so I was concerned about the second order distortion/output polarity effect. So once you find which speaker wires polarity works best, changing the input polarity will retain the good effect at the output? I figure it will but would appreciate your assurance.

So which is more likely to add goodness - less or more Vds?

Thanks for the advice.
 

Ben Mah

Member
Paid Member
2012-05-18 6:36 pm
Vancouver
Rick, I flip the speaker wires not to listen to the change in phase of recordings. I flip the speaker wires to change the phase of the second harmonic. With negative phase second harmonic the music soundstage is at the plane of the speaker and behind, and with positive phase second harmonic the music image projects out in front of the speaker.

The Singing Bush does not invert phase so the regular speaker connection should give you positive phase second harmonic, and flipping the speaker wires should give you negative phase second harmonic. Choice is a personal preference.

As for Vds, I have found that an increase in Vds increases second harmonic distortion. Whether you will like it or not will depend on your preference and the initial amount of distortion. The other consideration is that Tokin SITs have large variations in their properties. With your four channels, each channel may have different levels of distortion. If you are considering experimenting with operating points, it is probably best to baseline the channels so that you know what distortions you already have.

The best tool for this is a program such as Room Equalization Wizard (REW) or ARTA, in conjunction with a computer sound interface card:

https://www.diyaudio.com/community/threads/howto-distortion-measurements-with-rew.338511/

You can also see effects of Vds changes in real time if you have the amplifier hooked up for distortion measurement while you are making the changes.

Another way to see effects of changes is to simulate the circuit in a program such as LTSpice. Although the results may not be entirely accurate, the simulations do give a very good idea of what is happening, from frequency response to harmonic distortion to voltages and currents within the circuit. I started using it two years ago and I have learned a lot about various circuits by simulating them - very educational.

https://www.diyaudio.com/community/...ding-ltxvii-from-beginner-to-advanced.260627/
 
Last edited:
Rick, this may help visualize this H2 phase. I made this as a pictorial from Nelson's discussions of H2 phase in amplifiers and Vds adjustments.

H2 Phase Notes.jpg


In this case, it's example #3, Inverting Amplifier (or common source) with Vds adjusted below Vds where the H2 would null (if twiddling the Vds while watching spectrum of the amp output).

Positive phase H2 rises along with the positive going signal. So in absolute phase, when the speaker cone is moving towards you, the H2 is adding to the towards-you direction.

-H2 is the opposite. -H2 falls on the positive going signal. Speaker cone moving towards you, -H2 subtracts from the towards-you direction.

+H2 more intimate, warmer and what I call, in your face.
-H2 more spacious, deeper sound stage, sitting in audience sounding.

It's all personal preference. My examples end with -H2 because that's what I prefer.

One comment that I agree with Nelson on is that absolute phase is important. For instance a kick drum hit should push the speaker towards you not suck the speaker away, that would be weird.

Cinco
 
As usual being late to the party, I got some Tokin SITs (THF-51S) before all SITS disappeared from reliable sources. Impatiently waiting for Pa to reveal circuit for these Toking Biguns, I started trying to figure out what to do with them. Scouring diyA I considered all sorts of options (mostly from ZM and Ben Mah).

Follower seemed boring and a waste of this device, Active loading will be hot as hell (especially summer time in Texas). Hmm... leaning towards choke load.

Lately my favorite amp to listen to has been a Choke Loaded Lu'amp (basically Pa's Zen v8 with choke replacing the lightbulbs for load and half the power supply voltage). There is something just really likeable about this amp's organicity.

:idea: I'll just do a SIT version. This SIT Amp will end up being even simpler than that one.... so I thought.

More reading of threads later I realized I need a buffered bias voltage... Aha :scratch2:
I could use the lower half of Singing Bush and done... Then I read this:

View attachment 1041895

Ok so I need power sequencing. ON Bias Supply -> Drain Supply..... OFF Drain Supply -> Bias Supply
This is getting even more complex. I even considered a Power Switch + Standby Switch to keep things simple (like old solid state rectifier tube amps have so plate voltage does not hit power tubes before they warm up).

Mighty ZM to the rescue! Choked Singing Bush solves all problems. I was trying to reinvent the already ZM invented wheel (no wonder I'm a ZM fan :worship:).

I should have read the whole thread.

The big question for you, Zen Mod, is when will these boards become available?
No hurry, just wondering when this iron comes out of the Zen Mod fire... there are so many irons in that fire all at once. :rofl:

Cinco

yup, there is that thingie that SIT is Normally On, so you must take care to choke it first ( whichever way) then bias gradually
no problem with Singing Bush ( as there is ever any problems with Singing Bush - we all know that he's singing all the time, while not bothering to hear what we are questioning, but luckily he's always already having in hand more than we asked for) , small biasing supply is always faster to get On than main rail is - that even in case of having both secondaries ( small and big) on same Donut

now - Choked Singing Bush - just one of several I already have pcbs on shelf
while there are several needing just pcb confirmation, all development already done Ikebana style on T-Bed, Choked Singing Bush needs confirmation of everything, not being tried at all
main obstacle/stumbling point is Power Up behavior ; I don't care for SIT - small supply is fast enough, opto-CCS-whatever else biasing thingie is slow ramping, but choke in Drain is fast route for topping up output cap, and if speaker is firmly connected, there is no other route for (output cap) topping up current than through speaker itself
and that current is fast and big, at least judging with common sense and simulations

now, that need to be observed and taken care of ( if needed) in simplest possible way

besides fact that working on amp itself demands time, which I need to find, soon ........
 

Ben Mah

Member
Paid Member
2012-05-18 6:36 pm
Vancouver
Re: absolute phase.

I agree absolute phase is important. However it may not be easily achieved as my understanding is that most recorded material that is available to the consumer may be in or out of phase. Even individual songs within an album may have some songs in phase and some out of phase. So trying to adjust for that is not something that I attempt to do, or even attempt to determine whether a song is in or out of phase.

However, choosing negative or positive second harmonic distortion is something that I can do.
 
now - Choked Singing Bush - just one of several I already have pcbs on shelf
while there are several needing just pcb confirmation, all development already done Ikebana style on T-Bed, Choked Singing Bush needs confirmation of everything, not being tried at all
main obstacle/stumbling point is Power Up behavior ; I don't care for SIT - small supply is fast enough, opto-CCS-whatever else biasing thingie is slow ramping, but choke in Drain is fast route for topping up output cap, and if speaker is firmly connected, there is no other route for (output cap) topping up current than through speaker itself
and that current is fast and big, at least judging with common sense and simulations

now, that need to be observed and taken care of ( if needed) in simplest possible way

besides fact that working on amp itself demands time, which I need to find, soon ........
Yay, just waiting for ZM lab testing... great summer time project.

On the speaker charging output cap, I was planning to use Mark's Tuba muting relay circuit to short circuit speaker during power up/down.
 
Ben Mah, thanks for your advice.

As far as "absolute" phase I know it is impossible to know for sure. I should have added that more often than not there is a position that sounds best. That is what I settle on.

With my processor I can use an IPhone to switch polarity at the listening chair.

I have been working with REW for many years. I have long wondered how accurate beyond relative the distortion aspect of REW is. Nonetheless relative is better than nothing.

I figured the second order distortion aspect would simply be the speaker wiring and one could retain the ability to change polarity at the inputs for listening preferences without interfering with that.

This weekend I will see how either of the connections compare in REW distortion window.

It is amazing how low the distortion is with the CELESTION AxiPeriodic drivers (above 500 hz) are for the most part other than a couple of high spots. Less than 0.3 percent over most of the range around 93 dB average which seems almost hard to believe.

I guess as a starter I will take a closer look at the distortion of each channel's drivers and see what I see.

THANKS very much,
 
For some rainy fun I measured the distortion of either polarity with REW with my usual microphone set at ear level at my listening chair where all measurements are made.

I made numerous sweeps to make sure what was reported was real and it would be the same every time.

Second, third, fourth and fifth harmonic distortion were about half with one polarity all in about the same proportion to each other. This took me by surprise.

I am listening to the low distortion polarity for a few days and then I will switch back to all being in the, roughly double, distortion polarity. Figuring there might be some advantage to all being equal in this respect. There might be an advantage to this but it will take some time to erase the expectation bias of either polarity.

None of the amplifiers area dealing with capacitor or chokes other than the output coupling cap.

Ben Mah - how much more Vds do you surmise is needed for a real improvement? I have done nothing so far to change that,

I assume the higher Vgs should still retain the same Iq.

Whether my easy trial would correlate with measuring the amplifiers driving a dummy load is something that will have to wait for me to make the gadget to allow REW to measure the amplifiers alone.

I did think this was interesting in that the proportion of second to the higher harmonics did not change but all were lowered by about half.
 

Ben Mah

Member
Paid Member
2012-05-18 6:36 pm
Vancouver
Hi Rick, I don't have my THF-51S in common source mode any more as I converted it to common drain. However when it was in common source, I varied the Vds by about 2V. I also had 0.5R between Mosfet and SIT and took the speaker out from between 0.2R at Mosfet and 0.3R at SIT. More here: https://www.diyaudio.com/community/threads/the-singing-bush-tips-n-tricks.357497/post-6322682 There is an error in that post as I mentioned the speaker out was between 0.3R and 0.3R resistors. It was between 0.2R and 0.3R as mentioned above.

As for Iq, within a range of Vds the Iq does remain reasonably constant, but Iq does drop a bit as Vds rises.

I have not measured the speaker output distortion at my seating position, only amplifier out distortion. So in one position the amplifier distortion is out of phase relative to the speaker distortion and canceling and in the other position it is in phase and adding?
 
Thanks, Ben Mah.

I had corrected for input polarity when making the change at the speaker outputs but found the distortion measured the same for each polarity at the inputs. I had to see if this was a factor and in my case it was not.

Two volts is not much - I was worried you were speaking of a larger increase than that! Something to play around with the next rainy Saturday.

I know my measurements are only good for some relative comparison but being more than double made it a real difference.

I would imagine every speaker would respond differently so I did not bother to list which connection at the amplifier went to which at the speaker. Whether these differences would be nulled or exaggerated by coils and caps is something someone else will have to investigate.

Half of 0.4% (second) is probably inaudible. Most of the range is far lower than that. And with the lower distortion connection the majority is well below the noise floor with the Celestion Axi-periodic. Completely different case with the woofers pumpng through those holes in the side of the horn and the room becoming far more influential. Still the same trend is plain to see.

But down there there are places where the distortion reaches one percent and one or two blips of two percent.
 
Sad to see this thread has been relegated to page 8 so I will ask a question I have asked before - many years ago in regards to SIT 1. Not just to bring it forward - I am wanting to try this.

With capacitor coupled amplifiers could there be any value from using the series connection with a bias voltage applied to the "center" of the two capacitors?

How much does the output capacitor affect the output signal other than the important duty of blocking the DC? Is there any need to worry about this cap's affect on sound quality? Any advantage at all of making the output cap non-polarized? I figure with the back to back connection the bias is needed. One can use actual non-polar caps but those can take up lots of room if using the oil filled cans. That is what I am doing now.

Would nine volts be sufficient or should one consider using more? Like using a voltage divider from the main supply? At least you would not have to worry about the battery.

Of course, there are those who say even non-polarized caps benefit from this connection. I know JBL has done this in their crossovers but I do not know what caps they are using. Could get expensive with film caps.

I checked with MOUSER yesterday and it seems they have cancelled their order for the KEMET slit foil capacitors. They were on order in December and now they are special order. I had checked in between - they never received any of them. Thought that was interesting.

Any thoughts and advice is appreciated.