Hi cromodora,
-Chris
No, but DC across the control will make it sound scratchy. Include the servo.
No, they represent very low thermal noise. Just don't use cheap carbon resistors. You could use MF if you would like. Some transistors are noisier than others, so use transistors intended for low noise. They should be matched and connected thermally.
-Chris
Hi Chris,
semis, yes you get everything here but ironically I ordered from Digikey precisely bec. I didn't quite trust what I would get over here. Lots of gold plated brand named connectors in the stores here with rust on them if you see what I mean.
Off for the w/e, will come back on your post later.
semis, yes you get everything here but ironically I ordered from Digikey precisely bec. I didn't quite trust what I would get over here. Lots of gold plated brand named connectors in the stores here with rust on them if you see what I mean.
Off for the w/e, will come back on your post later.
Hi MBK,
The problem is that I want those Japanese numbers. Are they any trustworthy places there? Overstock from factories?
These are a very iffy proposition over here.
-Chris
The problem is that I want those Japanese numbers. Are they any trustworthy places there? Overstock from factories?
These are a very iffy proposition over here.
-Chris
I finished last night replacing the OT BJTs (U3,U4) with lateral mosfets (BUZ900,BUZ905) on my Symasym prototype boards. After making some simple modifications to the board and changing four resistor values, it worked great. No oscillations, thermally very stable.
Sound wise, fantastic. Does not lack in performance to the BJT version at all. THD wise, very similar, maybe even a slight improvement. At 80 watts into 8 ohms at 1Khz I measured a 2HD of -94db, with 3HD, 4HD, and beyond under -100db.
I was wondering, with only a single lateral mosfet for the OT per rail, I could eliminate the drivers (U5,U6), and replace the BD139 with a pot to control the bias since there's no need for thermal tracking. The rest of the circuit would remain the same. Any thoughts?
Sound wise, fantastic. Does not lack in performance to the BJT version at all. THD wise, very similar, maybe even a slight improvement. At 80 watts into 8 ohms at 1Khz I measured a 2HD of -94db, with 3HD, 4HD, and beyond under -100db.
I was wondering, with only a single lateral mosfet for the OT per rail, I could eliminate the drivers (U5,U6), and replace the BD139 with a pot to control the bias since there's no need for thermal tracking. The rest of the circuit would remain the same. Any thoughts?
Hi AAK,
I strongly feel that would be a big mistake. Mosfets should be driven from a low impedance source. All mosfet amps I've ever heard that sounded good had a real driver stage. Example - Luxman LV-105.
You can easily try that on another board and A/B them.
-Chris
Yes.
I strongly feel that would be a big mistake. Mosfets should be driven from a low impedance source. All mosfet amps I've ever heard that sounded good had a real driver stage. Example - Luxman LV-105.
You can easily try that on another board and A/B them.
-Chris
Hi anatech,
Thanks! I'll try it on another board and A/B them to quench my curiosity. I think using lateral mosfets on my version of Symasym may be a better choice than BJTs considering that my board doesn't allow for mounting the drivers on the main heatsink causing some thermal drift though minimal. It's also nice not having to mount the BD139 to one of the OT. Of course the laterals are more expensive.
Regards,
Al
Thanks! I'll try it on another board and A/B them to quench my curiosity. I think using lateral mosfets on my version of Symasym may be a better choice than BJTs considering that my board doesn't allow for mounting the drivers on the main heatsink causing some thermal drift though minimal. It's also nice not having to mount the BD139 to one of the OT. Of course the laterals are more expensive.
Regards,
Al
I wouldn't drive the mosfets without drivers unless you want a subwoofer amp... The current from vas is not sufficient to drive the mosfets at higher frequencies.
I also wouldn't skip vbe-multiplier, just don't mount it on the heatsink. You still need to compensate the drivers. Also, with a plain resistor you would loose bias stability as any change in vas bias would be amplified by the outputstage, worsening psrr.
my
Mike
I also wouldn't skip vbe-multiplier, just don't mount it on the heatsink. You still need to compensate the drivers. Also, with a plain resistor you would loose bias stability as any change in vas bias would be amplified by the outputstage, worsening psrr.
my
Mike
Hi Al,
Report back then. It's easy enough to try.
I agree completely with Mike on this, but it's a cheap and easy experiment.
-Chris
Report back then. It's easy enough to try.
I agree completely with Mike on this, but it's a cheap and easy experiment.
-Chris
Hi Chris,
back online...
Re: semis: tomorrow I'll be at Sim Lim Tower, "The Place" with dozens if not 100's of merchants. Will ask around for those numbers.
Re: scopes: what kind of work do used ones usually need? And hpow critical are the probes? I see generic probes for sale - good enough for beginners on a low end scope? I know from experience that having instruments is not enough, one also has to know how to use them, and if I end up buying one I don't intend to shoot too high and get nothing but artefacts...
Re: my own amps:
- I am using 2SC4793/2SA1837 for drivers and output 2SC5200/2SA1943, that's AAK's BOM and should already be pretty good no?
- startup hum/thump: I though it was more a matter of when the front end is "up" vs. the output. The 2N5551's Hfe is low but well within specs by data sheet and this is AAK's recommended BOM. Still a bit at a loss why I lucked out.
- front end: Say I find 2SA970, any other significant benefits than likely better startup behaviour? In theory THD might go down further but AAK's measurements are very good using the 2N5551, assuming mine are not to be trusted (10x higher THD than AAK) the numbers should be OK.
- other small signal: any benefits replacing all 2N5551/2N5401 with 2SA970/2SC2240?
I have to say I am very reluctant to desolder the small signal transistors on my 6 completed boards at all, I don't seem to be very gifted for desoldering or my iron is too weak, it is always a singed affair with occasional pushing and pulling with pliers and ugly noises. Very rare that I can suck out the solder cleanly with the standard braid I am using and the pumps I've tried seem completely useless.
back online...
Re: semis: tomorrow I'll be at Sim Lim Tower, "The Place" with dozens if not 100's of merchants. Will ask around for those numbers.
Re: scopes: what kind of work do used ones usually need? And hpow critical are the probes? I see generic probes for sale - good enough for beginners on a low end scope? I know from experience that having instruments is not enough, one also has to know how to use them, and if I end up buying one I don't intend to shoot too high and get nothing but artefacts...
Re: my own amps:
- I am using 2SC4793/2SA1837 for drivers and output 2SC5200/2SA1943, that's AAK's BOM and should already be pretty good no?
- startup hum/thump: I though it was more a matter of when the front end is "up" vs. the output. The 2N5551's Hfe is low but well within specs by data sheet and this is AAK's recommended BOM. Still a bit at a loss why I lucked out.
- front end: Say I find 2SA970, any other significant benefits than likely better startup behaviour? In theory THD might go down further but AAK's measurements are very good using the 2N5551, assuming mine are not to be trusted (10x higher THD than AAK) the numbers should be OK.
- other small signal: any benefits replacing all 2N5551/2N5401 with 2SA970/2SC2240?
I have to say I am very reluctant to desolder the small signal transistors on my 6 completed boards at all, I don't seem to be very gifted for desoldering or my iron is too weak, it is always a singed affair with occasional pushing and pulling with pliers and ugly noises. Very rare that I can suck out the solder cleanly with the standard braid I am using and the pumps I've tried seem completely useless.
Some notes on the sound (all compared to previous LM3886) though still no quiet time for critical listening:
- still got the impression of more detail, this seems to be real
- no grain
- but occasional pronounced sibilants
- bass same as LM3886 (but the LM3886 had the benefit of a separate XF for the bass with 22,000 uF caps. This didn't fit anymore with the Symasyms, so one XF for 3 amps @ 10,000 uF each)
- still got the impression of more detail, this seems to be real
- no grain
- but occasional pronounced sibilants
- bass same as LM3886 (but the LM3886 had the benefit of a separate XF for the bass with 22,000 uF caps. This didn't fit anymore with the Symasyms, so one XF for 3 amps @ 10,000 uF each)
For the record I built the simple RF probe. ! nF series, 1N4148 to ground, series 3Meg to DMM (it has 7.5 Meg input impedance and that should give an RMS reading when reading DC with the 3Meg series).
Result is 5.9 mV on amp, 5.7 mV shorted probe, so no reason to assume oscillation, but when I touch the probe I get 150 mV or so therefore I assume it does work 😉 . This was with the L // R output inductor on the amp.
Next if I find a Germanium diode I might get higher sensitivity, and an analog multimeter for real time burst measurements. How sophisticated 😀 .
Result is 5.9 mV on amp, 5.7 mV shorted probe, so no reason to assume oscillation, but when I touch the probe I get 150 mV or so therefore I assume it does work 😉 . This was with the L // R output inductor on the amp.
Next if I find a Germanium diode I might get higher sensitivity, and an analog multimeter for real time burst measurements. How sophisticated 😀 .
Hi MBK,
Glad to see you are back up and running.
I'd like to know how much those drivers (2SC4793/2SA1837) you are using go for. Since the 2SC2238 and 2SA968 went away, I've been very sad.
Try the 2SC2240 and 2SA970 out on a channel and see how you like it. This may not be a numbers thing, but rather a sound thing. Maybe no-thing! 😀 (sorry for that). If you like them, they would go well in midrange or HF channels. Full range also.
When buying a 'scope, you should always get good leads. Your signal has to go through these and it's no good if the signal is distorted or they break a lot. Tek probes are good, so are Agilent and even my Philips probes. Take good care of them and I expect them to last a long time.
A 100 MHz, dual channel model would be pretty basic in used stuff. A 20 MHz model would do for many things. You need 100 MHz to look at FM front end stuff and if makes for a clearer picture of 10 MHz signals (FM is 10.7 MHz, clocks are 10 MHz). If you go digital, your requirements go way up to 500 MHz or better to get a clear trace. Therefore, stay with an analog 'scope. Avoid Philips.
There are some better RF probes you can make. I rattled off a very basic one. Some look like a voltage doubler. At high impedances, the leakage of germanium diodes may pose a problem. I am not sure on this but I do know the leakage is higher.
-Chris
Glad to see you are back up and running.
I'd like to know how much those drivers (2SC4793/2SA1837) you are using go for. Since the 2SC2238 and 2SA968 went away, I've been very sad.
Try the 2SC2240 and 2SA970 out on a channel and see how you like it. This may not be a numbers thing, but rather a sound thing. Maybe no-thing! 😀 (sorry for that). If you like them, they would go well in midrange or HF channels. Full range also.
When buying a 'scope, you should always get good leads. Your signal has to go through these and it's no good if the signal is distorted or they break a lot. Tek probes are good, so are Agilent and even my Philips probes. Take good care of them and I expect them to last a long time.
A 100 MHz, dual channel model would be pretty basic in used stuff. A 20 MHz model would do for many things. You need 100 MHz to look at FM front end stuff and if makes for a clearer picture of 10 MHz signals (FM is 10.7 MHz, clocks are 10 MHz). If you go digital, your requirements go way up to 500 MHz or better to get a clear trace. Therefore, stay with an analog 'scope. Avoid Philips.
There are some better RF probes you can make. I rattled off a very basic one. Some look like a voltage doubler. At high impedances, the leakage of germanium diodes may pose a problem. I am not sure on this but I do know the leakage is higher.
-Chris
Hi Chris,
surprise, seems you are in luck. I found one dealer with plenty of Toshiba cartons labeled "1600 pcs" etc. Out of your list he claims he has the following items, prices based on purchase of 100 pcs. I converted the Singapore dollars to USD and CAD courtesy Oanda.com:
type SGD USD CAD
2SC3421 0.30 0.21 0.22
2SA1358 0.20 0.14 0.15
2SC3467 0.20 0.14 0.15
2SA1370 0.22 0.15 0.16
2SA1360 0.40 0.28 0.29
2SC4793 0.30 0.21 0.22
2SA1837 0.22 0.15 0.16
Ironically he does not have the 2SC2240 and 2SA970 for my amps
. But I did not have time to check all dealers there, surely someone will have them...
Bougth a $12 analog meter and some Germanium diodes for more experiments. But the meter is just as slow as the DMM and hardly reacts at turnon. The DMM on fixed scale setting gave a max of 2.5V but really just for a fraction of a second.
The more I think of it my amps are probably fine, just the high eff. driver makes the turnon thump appear massive.
surprise, seems you are in luck. I found one dealer with plenty of Toshiba cartons labeled "1600 pcs" etc. Out of your list he claims he has the following items, prices based on purchase of 100 pcs. I converted the Singapore dollars to USD and CAD courtesy Oanda.com:
type SGD USD CAD
2SC3421 0.30 0.21 0.22
2SA1358 0.20 0.14 0.15
2SC3467 0.20 0.14 0.15
2SA1370 0.22 0.15 0.16
2SA1360 0.40 0.28 0.29
2SC4793 0.30 0.21 0.22
2SA1837 0.22 0.15 0.16
Ironically he does not have the 2SC2240 and 2SA970 for my amps
. But I did not have time to check all dealers there, surely someone will have them...Bougth a $12 analog meter and some Germanium diodes for more experiments. But the meter is just as slow as the DMM and hardly reacts at turnon. The DMM on fixed scale setting gave a max of 2.5V but really just for a fraction of a second.
The more I think of it my amps are probably fine, just the high eff. driver makes the turnon thump appear massive.
Hi MBK,
Those appear to be very good prices. I'm going to have to check my finances. Will he sell in less than 100 quantities? Not 1sy 2sy, but in say 20 or 30 and 50 lots?
Also, I don't see 2SC3423 listed. Your 2SC2240 and 2SA970 are also ones I use a lot of.
Is there any way to get a list of what he carries? I could recommend some others for you. Meantime I'm going to think on this.
-Chris
Those appear to be very good prices. I'm going to have to check my finances. Will he sell in less than 100 quantities? Not 1sy 2sy, but in say 20 or 30 and 50 lots?
Also, I don't see 2SC3423 listed. Your 2SC2240 and 2SA970 are also ones I use a lot of.
Is there any way to get a list of what he carries? I could recommend some others for you. Meantime I'm going to think on this.
-Chris
The 3423 was not in stock there.
One other thing I noticed back home when I looked up the data sheets is that they also come in different hFE categories and that is one thing I did not ask him - which hFE range those are in.
Complete lists are a bit hard to get from those shops. Whenever you ask for a quote they act very suspicious because the local custom is to send mules to the competition to check on prices to undercut them. Prices are not marked anywhere, he pulls them out of his hat / the boss. I was surprised he agreed at all to write down my list at all on paper. So we'll be stuck to ask piece by piece. One of the drawbacks to buy from that kind of retailer here, razor thin margins mean, not exactly MBA refined customer service improvement plans.
Quantities: for small signals. e.g. 2SC2240 and 2SA970 I might want to get 50 each PNP and NPN myself, for matching you need some quantity anyway. The others I suppose he'll sell lesser quantities as well (making a face of course), for a higher price. Or you organize a GB for aficionados 😱
One other thing I noticed back home when I looked up the data sheets is that they also come in different hFE categories and that is one thing I did not ask him - which hFE range those are in.
Complete lists are a bit hard to get from those shops. Whenever you ask for a quote they act very suspicious because the local custom is to send mules to the competition to check on prices to undercut them. Prices are not marked anywhere, he pulls them out of his hat / the boss. I was surprised he agreed at all to write down my list at all on paper. So we'll be stuck to ask piece by piece. One of the drawbacks to buy from that kind of retailer here, razor thin margins mean, not exactly MBA refined customer service improvement plans.
Quantities: for small signals. e.g. 2SC2240 and 2SA970 I might want to get 50 each PNP and NPN myself, for matching you need some quantity anyway. The others I suppose he'll sell lesser quantities as well (making a face of course), for a higher price. Or you organize a GB for aficionados 😱
Just babbling on...
Finally made progress in bringing the grounding scheme in active x-o and preamp into line with the now earth-lifted scheme. I can still hear some minor noise pickup but even as it is now, it brought significant sonic benefits.
The amps now sound decidedly better than the LM3886, and up to expectations. More detail, especially treble, yet smooth and no more apparent sibilance. Probably as a result of this, subjectively the bass improved as well (bearing in mind that tamed highs mean relatively stronger bass). As usual, implementation and layout are crucial. Still some work to do but sonically I am very happy now.
Finally made progress in bringing the grounding scheme in active x-o and preamp into line with the now earth-lifted scheme. I can still hear some minor noise pickup but even as it is now, it brought significant sonic benefits.
The amps now sound decidedly better than the LM3886, and up to expectations. More detail, especially treble, yet smooth and no more apparent sibilance. Probably as a result of this, subjectively the bass improved as well (bearing in mind that tamed highs mean relatively stronger bass). As usual, implementation and layout are crucial. Still some work to do but sonically I am very happy now.
Thanks, I'm glad myself to finally be elated. The logic of my previous earth connected grounding actally worked well with the LM3886. When I lifted the signal ground for the Symasyms I had to make significant changes to the entire signal chain, and it took a while to get it right again.
Also found out, silly me, I had a bag of BC550C in the drawer. With my 38V rails I could have used those for input with a 5x higher hFE than the 2N551 and likely no nasty turnon thump. Ah well. When I have the time 🙄 .
Also found out, silly me, I had a bag of BC550C in the drawer. With my 38V rails I could have used those for input with a 5x higher hFE than the 2N551 and likely no nasty turnon thump. Ah well. When I have the time 🙄 .
does anyone know if there are anymore pcb's available for the mike b symasysm amp available. i have been reading the posts and it appears to be a good design. thanks
try,
try AAK Symasym PCB groupbuy that this build thread is about.
Ryssen is still taking orders and I feel like ordering some more.
MBK,
I share your sentiment about desoldering. I've blinded a few pads on this PCB and replacing parts is a hair pulling ordeal. I thought I have the neatest build before the thermal tracker (BD139) relocation saga. Also, replaced resistors and caps now reside underneath PCB. Its starting to look a mess and I feel like chucking it all and starting over with new boards to remotivate myself.
try AAK Symasym PCB groupbuy that this build thread is about.
Ryssen is still taking orders and I feel like ordering some more.
MBK,
I share your sentiment about desoldering. I've blinded a few pads on this PCB and replacing parts is a hair pulling ordeal. I thought I have the neatest build before the thermal tracker (BD139) relocation saga. Also, replaced resistors and caps now reside underneath PCB. Its starting to look a mess and I feel like chucking it all and starting over with new boards to remotivate myself.