UGS adventures

Re: Output Impedance

Hi Willus,

z_willus_d said:
Cheff, as for lowering the value Ratio on the resistors in my attempt at adding Tino’s BJT bias circuit, that doesn't work for me. With the current values I simulate ~4.5mA through the JFET and the Vgs is ~0V. I did note in the 2sk170 datasheet that a Vgs of -400mV gives a better/flatter Ids curve, but I wasn’t able to obtain the correct 10V/-10V split with that Vgs. Lowering the resistor value ratio for higher currents threw off the balance of the voltage divider.

Err... I was not speaking of Tino's circuit in this case... Sorry for
misunderstanding... In Tino's case, you'll have to get the b(i)ase
voltage of the folded cascodes close to the Jfets drain voltages
(just a Vbe apart). So adjust the CS and resistors to get it...

Nevertheless, isn't 4.5mA through the JFET nearly ideal? That should be more than enough to overwhelm the minor Base currents in the uA range, right? I agree that this CS adds little benefit other than PSRR, but it did seem to help balance the base-voltages, whereas with the simple resistor divider I get a slight imbalance of base bias voltage between the NPN and PNP transistor nodes.

Base voltages for the cascode on "my" version are not as much critical
as in Tino's. A slight imbalance can be lived with - IMHO-, it doesn't
have adverse effects on sound nor on offset.

Cheers,
 
Hi Cheff

CheffDeGaar said:
Hi Rob,

No, 3 volts won't hurt. You can even more decrease the supply voltage,
but you'll have to take care of the cascode biasing voltages to get
9-10V at the bases.

BTW, have you measured the DC offset ?

Cheers,

I can keep it below 1mv and its very stable it changed a little with different input voltages that's all.

Rob
 
Hi

Yesterday evening we did a shootout between the UGS Module from Cheff and a Balanced Tube preamp from myself.
I also did a shootout between the Tube Preamp and a Theta Casablanca III and a Six Shooter which I own.
I can tell that the tube amp wins in the midrange but misses a bit of air on the top end against the Theta.
I did some changes on the tube preamp and find some things in that area and have tot do the shootout again but I think it’s now in overall favour for the tube.
Back to the UGS Module in short it can’t beat the tube amp but it sounds very well no harshness or a typical transistor sound.
In detail nothing is wrong with the Module but it’s not so musical as the tube amp.
It’s very difficult to explain but with the tube amp you get goose bumps or how do you call this and with the UGS Module not .
The Soundstage is smaller and has less air than the Tube.
With the module is nothing wrong and you miss nothing till you change.
I think also that the tube amp is a little more dynamic and maybe we can change that by lowering the overall feedback so to have more gain.
Also I leave it on al day before listening and the offset was always below the 1mV.
So I think we can also increase the value from the 1k resistor to ground and listen what those 2 things bring.
I didn’t use output caps or input caps and I use a 50K alps potentiometer at the input not output.
The input was unbalanced and output balanced.
I listen to a Denon A1XV an aleph kind of amplifier and my own build speakers.
Maybe not what you want to hear but that’s it.

Rob
 

Attachments

  • shotall.jpg
    shotall.jpg
    96.5 KB · Views: 2,913
Official Court Jester
Joined 2003
Paid Member
Rob Dingen said:
Here is another shot from the 2 preamps side by side one is a bit empty but who cares.

Rob


in case that you have unbalanced input to ugs ,then entire virtues of susy weren't present

don't take me wrong,tube amps and preamps are my love too,but UGS is (at least how I see it) constructed for full balanced

for other combinations Papa have other more romantic things,if we chase romantic -goose bumps-as you say
I always do ;)
 
thanks for sharing!

looks like you used UGS in unbalanced mode?
any chance of trying balanced?

also curious if any way to try a version of UGS that has no overall global feedback? i'm curious about opinons on that ...

mlloyd1
edit: ooops, i see zen mod already commented on balanced and balanced operation for UGS ... :blush:
 
Hi
Yesterday evening we did a shootout between the UGS Module from Cheff and a Balanced Tube preamp from myself.
I also did a shootout between the Tube Preamp and a Theta Casablanca III and a Six Shooter which I own.
I can tell that the tube amp wins in the midrange but misses a bit of air on the top end against the Theta.
I did some changes on the tube preamp and find some things in that area and have tot do the shootout again but I think it’s now in overall favour for the tube.
Back to the UGS Module in short it can’t beat the tube amp but it sounds very well no harshness or a typical transistor sound.
In detail nothing is wrong with the Module but it’s not so musical as the tube amp.
It’s very difficult to explain but with the tube amp you get goose bumps or how do you call this and with the UGS Module not .
The Soundstage is smaller and has less air than the Tube.
With the module is nothing wrong and you miss nothing till you change.
I think also that the tube amp is a little more dynamic and maybe we can change that by lowering the overall feedback so to have more gain.

Hi Rob, thanks for reporting.

What's the gain of your tube pre, and the input sensivity of your amp ?
I confess I'm a bit puzzled about dynamics, since I didn't feel a lack of,
rather the opposite ;) , but I'm not as experienced (and equipped)
as you are.

It could be related to the bandwidth you measured. I'm not quite sure
the perfboard layout is optimal in terms of parasitic capacitance,
and thus BW limiting, and this could impact the dynamics by slowing
transients. I will try to grab a decent generator and try to measure
BW on my modules with pcb, or at least see the shape of squares
signals at 100kHz...

And the same goes for soundstaging and air... I felt the opposite,
but didn't compare with a tube pre, just my old BoSoZ and the
"no preamp" solution :)...

But when it comes to goose bumps, that's only a personal question
of culture, education and taste :) And that was exactly the opposite
when a friend of mine switched from a tube pre to the ugs ;)
He was much more thrilled by the ugs than by the tube...

Maybe not what you want to hear but that’s it.

No need to feel sorry ;) I never meant it was the ultimate preamp,
and it surely can be improved. But the overall impression I retain
from what you wrote is that you miss the perceived "magic" of the
tube pre,otherwise nothing's really wrong...
And it's very hard to compete with magic ;)

Anyway, warm thanks for taking the time to build and test it.
Just tell us how it goes with a higher Rout and a gain modification.

But for the moment, we're at deuce : amongst the two that have
built and listened to the ugs, one love it, the other not...
Can't conclude :)

Cheers
 
Hi

I don't have a balanced source sorry.

Cheff I don't hate it and nothing is missing.
But there is better out there.
You hear inmediatly the difference.
And like transistors net every tube amp is good.
The only reason to build this preamp and other things is to improve my system.
The gain from my tube pre 10x and bandtwidht the same as the UGS 750Khz.
So the bandwidht is not the problem but maybe the input capacitane from the Jfets I don't no.
But maybe we can get it better.
Build my preamp and compare it to yours to find out.

Rob
 
Hi Rob,

Do you have a schematic of your tube pre ? And I know music is not
numbers (although Bach could be the exception ;) ), but could you post
the same measurements results on your pre than you did on
the UGS (distortion, output Z, etc...) ?
Just to compare, not to compete :)

Cheers
 
Hi Cheff

Ok I measure everything with 2V RMS.
Bandwidht 650Khz.
Gain Unbal in Unbal out 5x.
Gain Unball in Bal out 10x.
output imp 80 Ohm (but when the amplitude increase the output imp increase).
THD 0.02% one side not balanced and only 2e harmonics.
If you can make time and want to spend some money on parts build the tube preamp I think you would like it.
And no music are not numbers but emotion and one amp has it and the other not ;)

Rob
 

Attachments

  • tubepre.png
    tubepre.png
    24.1 KB · Views: 2,921
Very interesting!

I shall report my listening test soon but it will be Oct 8 as I am too busy these few days.

I only feel 50k volume is a bit high for the UGS.

Rob, I shall also build your pre-amp for comparison in my my system. Thanks for sharing the design and it look like a NP Aleph P plus a source follower. Yes we need your power supply schematic also.

Valve amp is tasteful and what output cap you use? Do you use any expensive resistor also?

Cheers,
Spencer
 
Rob, what kind of power supply design are you using for your UGS module? – what your pic shows seems a bit emaciated. Cheff was using isolated dual-mono power supplies with nice regulation and capacitance, and yours looks more like a linear regulator with input from a shared small, laminated core transformer. Your tub pre looked to have a more developed supply. Not that this is everything (I don’t know), but it could explain the lack of dynamics or, at least, separation. Can you post the supply schematic for your tub design to help in the compare? And what transformer are you using? Oh, and what make of tubes do you use?

I think THD should be lower for the UGS if the volume attenuation is at its input, while it may be more susceptible to noise before the JFET input. I noted Cheff had his attenuation at the output, a minor difference.

Gain Unbal in Unbal out 5x.
Gain Unball in Bal out 10x

Did you short the negative output to GND for the unbal out case?

While any comparison is fair, and desired, I wonder if comparing a tub pre to solid-state is a bit like comparing apples to pears – or some other crunchy fruit. I’d be interested in a comparison of the UGS with other “top-class” solid state amps. Were the highs lacking in clarity and “seamlessness” on the tub pre as compared to UGS?
 
Hi

I attach the schematics of the power supply but there is one modification.
Between the 7806 GND pin and GND is a diode the Bat43.

It’s possible that the 50k pot is the problem indeed.
I’m building a 7 bit ladder volume at the moment and the print boards are ordered so I have to wait.
I build the same power supply as Cheff with maybe a small difference in capacitane see the picture on the previous pages.
I did short the negative input to ground.
Don’t think this comparison is apples to pears they both are a part of a music system and I don’t care what’s inside of it.
If you read my story before the Theta Casablanca with SixShooter sounded better on a couple of things as the Tube Preamp and it sounds better than the UGS module for sure.
But like the UGS and the Tube Pre are both in their development stage and for both can made changes that maybe bring some improvement to the sound.
Like I said lower feedback and increase output resistor to ground and for the Tube I want to make a current source for the followers.
The Tubes I used are Philips and the Caps at the moment from Audin Cap.

Rob
 

Attachments

  • psupply.png
    psupply.png
    23.4 KB · Views: 2,676