USSA-5 Build with Review

Administrator
Joined 2004
Paid Member
Hi Gaborbela,
With matching transistors of any kind, it is critical they remain at the same temperature from transistor to transistor. You can use the hFE on a DVM to get a rough match (divided into groups). If you build a jig to do it right, you will be disappointed as to how many you thought were matched that are not very good matches. Let each transistor settle to room temperature, handle them with tweezers and do not even breathe on any of them. Yes, they are that sensitive.

The jig places the transistors in a long-tailed pair (differential pair) configuration and in physical contact with each other. Foam goes over top to isolate them a little, then a larger cover (box, pan, whatever) to kill air currents. You can easily achieve sub 1% matches doing it this way. And yes, it makes a difference.

A few members made some PCBs for them and I have checked mine and one other. It worked very well. One uses a fan. That one doesn't work as well, not recommended.
The tail current is adjustable and it's just a red LED - transistor current source. Nothing too fancy. The LED (3mm) pokes through the mounting hole on a TO-126 transistor (pass device) so that the temperature tracks between them. I used a 2SC3421 and a 2SA1358 for the PNP part. There are details in one of the Adcom threads and in a couple other places. I gave the design to the members here, so feel free to build one. If I can find the details, I'll send them to you. I lost the computer they were on (Ransomeware - nasty, nasty, nasty).

-Chris
 
Storm in a teacup

Hi Fab,

So, my heatsinks turned up and I've proceeded with the USSA-5 assembly. At the moment I am up to the testing of the Front End where I must set 0.3VDC between test points using a couple of potentiometers. The first amp worked as you would expect with 0.3VDC found quickly and it being stable in both channels.

However, the second amp has an issue where the 0.3VDC does not stay stable. I set it using the first pot, then set it in the second pot and go back to check the first and it the voltage has moved. So I reset it to 0.3VDC and check the second pot and now it has moved...and round and round I go ultimately getting nowhere.

I've looked at the components and both boards have the caps and diodes in the same orientation and the driver transistors are in their correct positions.
I'm not sure what to check next. My guess is that one of Q1, Q2, Q3, Q4 needs replacing...but that is only a guess.

Cheers,

Anthony

So I went back to the amp that "works" and redid the setting of the 0.3vdc. It turns out that the first time around I set one pot to give the right voltage and the second was already correctly set and did not require any adjustment thus setting an expectation for an easy result in the second amp which did not eventuate. Anyway, once I shifted those pots in the good amp it behaved exactly as the problem amp. A second DMM and it all came together quickly for both amps so the build progresses... hopefully I'll get everything done tonight so that I can order the final resistors.
 
Yes, both at the same time and at the same rate.

Given the grief the heatsinks have given others earlier in this thread I decided to pull out my thermal imaging camera for my smartphone and check on the first amp 2.5 hours into warmup running the 50mVDC as set in the manual. The big heatsinks are barely lukewarm.


Seems to me that the right things are warm or not so warm. The fin side of the heatsink shows two warm patches where the heat radiates at the output FET's with gradually less radiation further from these points. Likewise for the power supply the diodes and resistors are a little warm and the first filter caps are doing more heavy lifting than the second pair. The toroid is also a little warm, but not exceedingly so and the soft start board has not got much going on at all, which is good.

Ambient temp in the room is a chilly 23C/73F (I've got a jumper on)
Output FETs both 47C/116F (measured in centre of their case)
Heatsink adjacent to the output FET's 40C/104F
Diodes in power supply 30C/86F
Warm resistors in power supply 32C/89F

Looks good so far.
 

Attachments

  • flir_20181123T105221.jpg
    flir_20181123T105221.jpg
    130.3 KB · Views: 330
  • flir_20181123T105310.jpg
    flir_20181123T105310.jpg
    922.9 KB · Views: 303
  • flir_20181123T105334.jpg
    flir_20181123T105334.jpg
    123.4 KB · Views: 320
  • flir_20181123T105404.jpg
    flir_20181123T105404.jpg
    173.4 KB · Views: 320
I did not know smartphone camera could do thermal imaging...

Fab

This is what I have.

There was a stupid special at a local supplier (just when I was looking at possibly getting one) which brought it into my purchase price comfort zone...just. Nonetheless it is very handy.

On another point, I've done the calculations for the "Output Mosfet VGS Determination" of the first amp and I have a negative value for RV2. The manual I have here is not clear on how to do the calculations so perhaps I've made a mistake...earlier in this thread someone else posted formulae which are not in my manual so I've used them.

Measured VGSN = 0.9600VDC (still moving about a bit after a few hours warming up, but my Fluke DMM measures to three decimal places of a mV so I've rounded the reading)

Measured VGSP = 1.533VDC

RM3 = 1.533 / 0.02 = 76.65 ohm
RV1 = RM3 - 51 = 25.65 ohm (I used 51R instead of 47R for RV2B)
1/RV1A = 1/RV1 - 1/100 therefore RV1A = 34.5 ohm


RM4 = 0.96/0.02 = 48R
RV2 = 48 - 51 = -3 ohm (I used 51R instead of 47R for RV2B)

Does this seem correct Fab?


Thanks,

Anthony
 

fab

Member
Joined 2004
Paid Member
Hi Anthony
Thanks for the link.

Yes you are correct in your calculations and it follows the manual.

You have quite different VGS values for channels N and P. You have to reduce your 51 ohms to 47 ohms and the recalculate new target current and readjust for other polarity as well.
0,096mv/47ohms = new target current (instead of 20ma) for both polarities.
You can also simply use 51 ohms if you intend to use higher bias current in the end.

Fab

After adjusting the resistirs value , the goal is that with the final output bias, the driver current will be in the range of 22 to 25 ma (the latter the better for enhanced linearity of mosget drivers). The driver current increases when you adjust the pots to increase bias current. If you intend to use only 1 amp bias for final value then select 25ma driver current for resistors value calculation.

Fab
 
Thanks Fab, but now I am confused.

At the moment nothing makes sense for me...single ended valve amps were never this finnicky and I still don't really know what I need to do to get 1.3A bias and zero DC Offset. I've never built solid state or push pull before let alone required three multimeters to set the bias and DC offset.

I was hoping to shoot for a hot 1.3A bias because I have the cooling and power supply capacity, and those few extra watts may be useful, so this means 65mV between TP2 and V+, TP4 and V-. I read it that you are suggesting that the 51R RV2B may be ok at 1.3A bias?

I've not yet figured out how to measure the DC offset, so my VGS numbers could well not be right. I know what DC Offset is but I don't know where the second probe needs to be placed on the pcb so it can be measured...I've tried a couple of places but they both gave me different answers and the offset jumped about erratically...are you able to help me out there?

Sorry, but it seems I need a little hand-holding to get this amp finished. My window for getting it sorted is rapidly closing.

So the things I would like help with are:
(a) How to measure DC Offset at output?
(b)When determining the output mosfet VGS should I use 65mvdc instead of 50mvdc considering my 51R RV2B and my desire for 1.3A bias?

Thanks,

Anthony
 
Hey Greg, sorry didn't want to still your thread but didn't see the point to make another one as well just to post my review :D.

Ok before I start, I was waiting to listen to this amp through my new speakers set but that may take a while and I couldn't wait anymore, my current setup is a DSP controlled WAW (FAST), with SBA woofers and ScanSpeak 10Fs on top and crossed at 450Hz w Harsch time aligned XO. the signal goes through my AKSA Lender preamp before going through the DSP, the SBAs are powered by B&O Icepower and the 10Fs are running by USSA-5 now, so yeah this will not do the justice but it's what I have at the moment.

I m listening to this amp for about two weeks now, bias is stable at 49mv, with 1-3mv DC offset, the DF is about 26-27.
Right at the beginning you will notice the signature, it's there but subtle, not over pronounced. before I start going through some details if I want to describe this amp in a short sentence it would be "I couldn't wait to listen to next track and see how it would sound on USSA-5 !"

I have been listening to a wide range of music, from vocals of Vanessa Fernandez, crazy dynamics of Keith Jarrett music, Pink Floyd, Led Zeppelin, Neil Young, beautiful music of "If These Trees Could Talk"'s "The Bones of a Dying World" album and all the way to some old Metallica and crazy "The Aristocrats" and some Tom Waits in between.

the Sound:

The first thing I noticed was that the amp sounds very detailed, a lot of micro details, you will probably notice some stuff you haven't heard before and a huge sound stage, A unique signature is there but again, it's subtle and humble, One thing I keep noticing is that how the instruments are re produced and how detailed they are, it is like the amp reproduced them with more details and more attention and care, you can here every note clearly.

The Neil Young's 1975 Roxy - Tonight's The Night Live is a very good recording and it just sounds very good on this amp it has a new depth, the guitars and harmonica really has a 70s vibe to it.
The Roger Waters voice in "Paranoid Eyes" is very live like.

Recently I came across this track:
The Lords of the Secrets: Tasnif on Rumi Sonnets
It's Persian music if you like the eastern stuff, the recording is not the best as it has some ring to it but I like the first half of the song, A LOT of micro details with dynamics all over the place, it became one of my reference tracks, and it just sounds really fantastic on this amp, I really enjoyed it, check it out if you like.

I can go on about the different things that I like about this amp and get into more details but I rather leave it for after I listen to this amp through my new speakers.

The only down side about this amp is ... joking no down side, it's just that if the music is not well recorded this amp is ruthless, it WILL bring out the faults and the weakness.

Fab, I appreciate that you shared your design with DIY community, sorry for taking me so long to write a review, I still owe you some pics :D, hopefully after my new speakers.
 
Last edited:
Measured VGSN = 0.9600VDC (still moving about a bit after a few hours warming up, but my Fluke DMM measures to three decimal places of a mV so I've rounded the reading)

Measured VGSP = 1.533VDC

RM3 = 1.533 / 0.02 = 76.65 ohm
RV1 = RM3 - 51 = 25.65 ohm (I used 51R instead of 47R for RV2B)
1/RV1A = 1/RV1 - 1/100 therefore RV1A = 34.5 ohm


RM4 = 0.96/0.02 = 48R
RV2 = 48 - 51 = -3 ohm (I used 51R instead of 47R for RV2B)

Fab, I have looked at the schematics and think I have made a mistake in my calculations above even though you have deemed them ok. The RV2B as 51R instead of 47R is confusing because 47R is used in several parts of the calculations.

IM4pre = VGSN/(R14+RV2B) = 0.96/98 = 0.0097959 amps
RM4=VGSN/0.02 amp = 48 ohm
RV2 = RV2A & RV2B in parallel = RM4 - R14 = 48 - 47 = 1 ohm (so RV2A and RV2B in parallel must equal 1 ohm).
1/RV2A = 1/RV2 - 1/RV2B therefore RV2A = 1.02R because RV2B is 51R in my case, not 47R as specified.

I am quite sure that is the correct way to do the calculations. Sorry to mess you about Fab, but please could you confirm or deny my process.

Anthony
 

fab

Member
Joined 2004
Paid Member
Hi Anthony
Yes your calculation is correct. Anyway 1 or 2 ohms difference is not important.

(a) How to measure DC Offset at output?
Probes between “Out” and Gnd.

(b)When determining the output mosfet VGS should I use 65mvdc instead of 50mvdc considering my 51R RV2B and my desire for 1.3A bias?
No , you use 50mv to determine the resistor values as per your calculation correctly done. After that, increasing the output bias current to 1.3A will automatically increase about equally in the driver stage for both polarities.

Fab
 

fab

Member
Joined 2004
Paid Member
Hey Greg, sorry didn't want to still your thread but didn't see the point to make another one as well just to post my review :D. <snip>

Hi Aatto
You do not have to be sorry about this nice review of the amp :wave2:
Thanks !
I understand that you have not used the USSA5 for the bass yet.
Looking forward to hearing from you when you try it with your new speakers and hopefully driving also the woofer so you can evaluate the bass response. Of course it will not be as tight as a powerful class D amp but it may have some interesting characteristics too..


Fab
 
Last edited:
Done!

Hi Anthony
Yes your calculation is correct. Anyway 1 or 2 ohms difference is not important.

(a) How to measure DC Offset at output?
Probes between “Out” and Gnd.

(b)When determining the output mosfet VGS should I use 65mvdc instead of 50mvdc considering my 51R RV2B and my desire for 1.3A bias?
No , you use 50mv to determine the resistor values as per your calculation correctly done. After that, increasing the output bias current to 1.3A will automatically increase about equally in the driver stage for both polarities.

Fab

Thanks Fab,

Both amps are now tested and measured. VGS were almost identical on both amps: Amp1 VGSN = 0.96vdc and VGSP = 1.53vdc; Amp2 VGSN = 0.95vdc and VGSP 1.53vdc. This gives me resistors to order for Amp1 0.5R/42.2R and Amp2 1R/42.2R. That is for 20mA.

If I use 20.2mA then I could use a piece of wire and 40.2R in Amp1 and 0.5R/42.2R for Amp2.

Resistors will be ordered tomorrow, then there will be a weeks wait for them to arrive. That will give me time to design the case.

Anthony
 

fab

Member
Joined 2004
Paid Member
Yes you can use standard resistors value (including piece of wire) with a different target current value than 20ma. As long as it is at least 20ma. You do not need to have exact same current for both polarities (about <10% difference is ok). Ensure your VGS values were taken when the amp had less than about 10mv DC offset or so (manual says 5mv).
Fab
 
Last edited:
from mannual rev 0.4.5 "´8.14 Front End amplifier testing
· Connect V+ and V- and PGND (the one close to V-) to power supply (from 15Vdc to 30Vdc).
· Adjust P1 for about 0,3Vdc between TP6 and V-.
· Adjust P2 for about 0,3Vdc between TP5 and V+."
Are P1 and P2 correct? I can change voltage on TP6 only with P2 and P1 with TP5. Minimun voltage I can get is 0,6V.

The problem was some resistors were missing. I forgot RV1B (100R) and RV2B (47R). R15 and R16 are not mentioned in the manual when they go into the board.
Now, in one board, I have 0,30V and 0,355V but in the other one I can't take the voltage to the target I have almost 0,4V difference between them. I tried RV2B 36R and 68R but the measurements were the same.
 

fab

Member
Joined 2004
Paid Member
Hi
See manual extract:
“8.1 Small resistors mounting
Mount all 0.5w resistors except : RV1A and RV2A (to be installed later after measurement)
For RV1B use 100 ohms and for RV2B use 47 ohms (as initial values before measurement)
Mount D1, D2, D3 and D4. Ensure to not mix zener and BAV21 diodes and respect pcb orientation”


RV1B and RV2B values have nothing to do with the measurement you are doing in section 8.14. Only P1 and P2 have an effect. These 2 pots are adjusted the same way as in the Pass F5 amp (in case if you need more details than specified).
Of course the steps 8.7 and 8.8.5 must have already passed successfully.
One thing to pay attention is the ground connection depending on the pcb version you have (refer to previous post on this thread). You can connect both connections on the pcb to the PSU if you are not sure.
Fab
 
Coming out

Hello Folks,

today I have my big coming out. Yes I am one of them too, who build an USSA-5! :snare::eek: And it is my first attempt when I try to write with pictures. Therefore lets cross my fingers.

After reading the review from ggetzoff I was very curious if could build it and if I will get a set of PCBs from Fab too. Luckily I annoyed him enough to get a set. Sorry Fab. I could not resist your excellent work and thank you for making that possible. :)

I collected all the necessary parts over the last months, renovating our house let me not as much time as I wanted. My first reviewer of the case did not wait long.

IMG_6776.jpg
That is what our cat is thinking about my project. It seams that she knows that it will be a nice and warm place when it is finished,;)

Here are the first pictures of soldering the USSA-5 PCB. The power supply, soft start and speaker protection PCBs are already finished. Thank you Project16 for making the PCBs and your kindness.

IMG_0316.jpg

Dreaming of an almost ready amplifier!

IMG_0317.jpg

Trying to fit all parts into the case. In the theory the case was much bigger. Where is all the place gone?

IMG_6844.jpg

Drilling the holes into the heat sink, always frightened to see the drill on the other side. ;)

IMG_6845.jpg

I hope I can animate a few other person to start the build or post their picture too. I will keep you updated on the progress.

Have fun and all the best. André

P.S.: Pinnocchio thank you for your support too. P.P.S.: I want to have the pictures right in place, but I do not know how to make this. Sorry!