Still a work in progress but is now located in Forum Problems where it can be linked to whenever needed.
http://www.diyaudio.com/forums/forum-problems/309064-how-attach-images-your-posts.html
I got it, problem it was be solved, like thimios says, i make secon zobel xonstructio with 10 ohm and 100 nf, and now it can be setting bias at 24 mv. In fact the problem is 4.7 ohm its not same colour with value , that resistor value is 1 ohm , no i uses 10 ohm, thank everybody ... its really good sound cool ampli....regard
6ohms error could be from an ohmeter error and/or from a resistor error.
6ohms in 150ohms is 4%. That is a big error, far beyond what a 1% resistor tolerance and a 0.5% DMM tolerance would add up to.
I have attached my simulation file with a link to the Cordell models I used.
Running this will show how well it copes with a reactive load. Various load capacitor values and signal input levels should be tried.
The small overshoots with 100n deteriorate when higher values are substituted.
Looking at the Vas section the benefits of pole-splitting are conferred by the Miller capacitor stabilizing Q5 in the other half of the differential pair there are 47k resistors with 330pF which load the collectors of Q6 and Q9 to ground.
Since there is a signal transfer between the circuit halves via the emitters of Q4 and Q5 there is a cross feed in phase opposition.
Since Miller capacitance in transistors varies with collector voltage (which is not constant) there is a question of equality of capacitance in either arm.
A 1uF capacitor in parallel with 8R with a square wave input will show how close the relationship is. The amount of ripple at high signal inputs shows this to be a troubled one.
The capacitors in the simulation are notionally perfect in terms of value matching and transparency to signals and stray circuit capacitance has not been allowed for.
Running this will show how well it copes with a reactive load. Various load capacitor values and signal input levels should be tried.
The small overshoots with 100n deteriorate when higher values are substituted.
Looking at the Vas section the benefits of pole-splitting are conferred by the Miller capacitor stabilizing Q5 in the other half of the differential pair there are 47k resistors with 330pF which load the collectors of Q6 and Q9 to ground.
Since there is a signal transfer between the circuit halves via the emitters of Q4 and Q5 there is a cross feed in phase opposition.
Since Miller capacitance in transistors varies with collector voltage (which is not constant) there is a question of equality of capacitance in either arm.
A 1uF capacitor in parallel with 8R with a square wave input will show how close the relationship is. The amount of ripple at high signal inputs shows this to be a troubled one.
The capacitors in the simulation are notionally perfect in terms of value matching and transparency to signals and stray circuit capacitance has not been allowed for.
Last edited:
I posted my .asc to help investigation of Mike Bittner's circuit by whoever may be interested.
This bears some resemblance to a project amplifier that appeared in the Australian Silicon Chip magazine in 2003 as SC480 which Altronics in that country as a kit is still available.
The circuit for this is attached.
Silicon Chip published a replacement for the SC480 earlier this year - this has better specs.
The LTP Vas has been dispensed with - in IC manufacture where settling time is quoted as an important parameter it is worth looking at the circuit representation to see what that indicates about the preferred Vas structure.
This bears some resemblance to a project amplifier that appeared in the Australian Silicon Chip magazine in 2003 as SC480 which Altronics in that country as a kit is still available.
The circuit for this is attached.
Silicon Chip published a replacement for the SC480 earlier this year - this has better specs.
The LTP Vas has been dispensed with - in IC manufacture where settling time is quoted as an important parameter it is worth looking at the circuit representation to see what that indicates about the preferred Vas structure.
Quite a few views - I have felt a heel for not having any form of solution to the snags I pointed out. I have now had a bit of time to look at the circuit more closely over the weekend and made some basic changes that might be of interest.
Normally I would run simulations with a range of capacitor values from 4nF up to 2uF over a range of signal input levels - The amended circuit can cope with 2uF and some other values however it could still do with further refinement. I am leaving that as a digest for others.
Normally I would run simulations with a range of capacitor values from 4nF up to 2uF over a range of signal input levels - The amended circuit can cope with 2uF and some other values however it could still do with further refinement. I am leaving that as a digest for others.
For those who prefer lead compensation, C4 in the latest simulation could be moved to that position in parallel with R29.
Running the simulation for a variety of capacitor values and signal levels should be investigated for satisfaction.
If the values of R12 or R15 are altered, the value of R11 needs to be altered to be identical in value.
Hopefully most will have already seen from viewing the SC480 circuit that equal values are necessary to set Q5 and Q6 emitter balance point.
Running the simulation for a variety of capacitor values and signal levels should be investigated for satisfaction.
If the values of R12 or R15 are altered, the value of R11 needs to be altered to be identical in value.
Hopefully most will have already seen from viewing the SC480 circuit that equal values are necessary to set Q5 and Q6 emitter balance point.
Symasym as replacement for Ulti-Amp
Hello,
I own a pair of the old Ulti-Amp power amps, described in Elektor 5/94 and 6/94. Unfortunately, one of them is broken and it seems like I'm not able to fix it
Now, as there are cheap and assembled Symasym boards available on eBay, I was wondering if I could do the following:
Just replace the main amp boards of the Ulti Amps by the Symasym boards and keep the power supply and protection circuits of the Ulti Amp.
I think, it would not be a big effort to adapt the circuits, and, at the same time, save me a lot of time But I would like to have opinions on this, if possible.
Anyone interested?
P.S.
PM me if you don't have the schematics of the Ulti Amp.
Hello,
I own a pair of the old Ulti-Amp power amps, described in Elektor 5/94 and 6/94. Unfortunately, one of them is broken and it seems like I'm not able to fix it
Now, as there are cheap and assembled Symasym boards available on eBay, I was wondering if I could do the following:
Just replace the main amp boards of the Ulti Amps by the Symasym boards and keep the power supply and protection circuits of the Ulti Amp.
I think, it would not be a big effort to adapt the circuits, and, at the same time, save me a lot of time
But I would like to have opinions on this, if possible.Anyone interested?
P.S.
PM me if you don't have the schematics of the Ulti Amp.
Hi fvzeppelin,
Boy, are those boards ever cheap!!! Buy the pair that says it is good for 100W or higher. They aren't, but they use the 2SC5200 outputs. As long as they aren't fakes, you're good to go.
If it were me, I would buy the bare boards which then allows me to use the On-Semi devices and to match the input transistors. You can also up the quality of the parts (and I don't mean like in Audiophool land) and end up with something very special.
Sounds like you have a solid plan. Just don't believe they are good for 150 watts or higher. With two pairs of outputs, you are looking at a 100 watt board and at your supply voltage, you will be under that at 8 ohms. Well over into a 4 ohm load.
-Chris
Boy, are those boards ever cheap!!! Buy the pair that says it is good for 100W or higher. They aren't, but they use the 2SC5200 outputs. As long as they aren't fakes, you're good to go.
If it were me, I would buy the bare boards which then allows me to use the On-Semi devices and to match the input transistors. You can also up the quality of the parts (and I don't mean like in Audiophool land) and end up with something very special.
Sounds like you have a solid plan. Just don't believe they are good for 150 watts or higher. With two pairs of outputs, you are looking at a 100 watt board and at your supply voltage, you will be under that at 8 ohms. Well over into a 4 ohm load.
-Chris
Hi Chris,
I was thinking of one of these:
One pair Assembeld Classic Symasym5-3 Discrete Power amplifier board 200W + 200W | eBay
In the driver stage, they use a pair of 2SC4793/2SA1837 and two pairs of NJW030/NJW0281 in the power stage (the ones you mentioned?). Don't you expect this to deliver by far more than 100W into 4Ohms? Don't know if they will deliver 200W burst - probably not
Yes, they are incredibly cheap, so I'm actually thinking of buying 2 PCBs plus twice the assembled version - just in case
They say, the transistors are coupled and DC offset is below 10mV for the assembled boards.But I actually don't really know how to place the Ulti-Amp's protections circuits in there. The DC offset shouldn't be a problem to sense, neither the temperature sensing, but I feel very uncomfortable with 2 topics:
- over current protection: don't know how to calculate part dimensions (T22 in third attachment)
- in the Ulti-Amp, they use opto couplers to switch off the output of the voltage stage in case of failures or thermal shut down. I'm not sure where to place the 4N35 in the symasym. A schematic of the eBay version is attached. Maybe place them between Q11/R17 and Q12/R21? I put the original version in the second attachment
I just uncovered a pair of these (Symasym 5_3) that I built about a decade ago—and then never listened too! A horrible sin. They got boxed up when I got married, and then a PSU never materialized. Anyway, I was excited to (finally!) listen to them, put them on the bench, powered them up, and r26 (33ohm) started smoking like a chimney. Pulled the transistors all of which tested ok. Looked for shorts, etc. etc. I seem to recall someone on this thread having some smoking resistors way back when. I'm on the hunt through this thread right now—but while I'm sifting through 400 messages, and chance someone might have an idea what would cause r26 to go up in smoke? Thanks kindly.
Oh lord. Pulled all the transistors and discovered I'd populated one board with two pairs of BC546s and no MPSA18s. Not sure of that's what caused the smoking resistor or if it was just my bad soldering. The other board was definitely just bad workmanship since cleaning and resoldering seemed to fix everything.
Anyway. Can I get these past ten years back?! Holy wow. Threw on Ratt's Out of the Cellar (probably not what these were intended for?) and the guitar hurts so good. Will have to spin some Luigi Nono later so that they really have something to chew on.
Thanks audiofan for chiming in on this. Thanks to the forum, and thanks especially to MikeB.
Anyway. Can I get these past ten years back?! Holy wow. Threw on Ratt's Out of the Cellar (probably not what these were intended for?) and the guitar hurts so good. Will have to spin some Luigi Nono later so that they really have something to chew on.
Thanks audiofan for chiming in on this. Thanks to the forum, and thanks especially to MikeB.