peranders said:
Yep, it is! 😀
With Rod keeping the schematic (details) for himself and Anthony taking a long, uncertain break etc., Lars' approach is the best thing that recently happened to DIYers looking for a tough class A/B MOSFET design with a public documentation of it's stability (even in a loose prototype, and thus showing easy repeatability). 😎
Sebastian.
Lars Clausen said:
the reason why I use carbons all the time. 😀
I have finished the PCB (I made four) for the zeta. it has everything on board except for the output stages, source resistors and gate resistors.
so if I need a low power amp, the only thing mounted off-board is a pair of mosfets. 🙂 the layout is basically similar to the schematic posted in the lcaudio website.
the plan is to use metal films on everything on the PCB. then, output stage gate resistors will be only carbon films. source resistors will be wirewound. will that affect the sound?
what about this......
local store does not list 2N5550, MPSA42 and MPSA92. but they sell 2N5551, MPSA43 and MPSA93. not sure if they sell them today. I have their old product list. will those work or do you have other recommended devices?
DjQUAN: Except than MPSA93 is only 200V, while MPSA92 is 300V, so they have a bit of safety margin. But if you don't use until +/- 100V it should be ok.
2N5551 ok though.
🙂
2N5551 ok though.
🙂
I plan to use two for +/-70V rails, and the other pair for +/-50V rails so I guess they're all fine.
does that mean the circuit (with the recommended devices ) can operate at higher than +/-100V?
(I don't plan to operate them at those voltages...just curious😀 )
does that mean the circuit (with the recommended devices ) can operate at higher than +/-100V?
(I don't plan to operate them at those voltages...just curious😀 )Could anybody suggest which is better of the following ;-
1, a class AB Nchannel amp with high supply voltage and low current ,
2, another amp with two class AB Nchannel amps in bridge mode having the same number of output devices as the above amp but here in two amps and having the same power output rating as the first amp , in this case the supply voltage will be lower with high current .
I feel that the first one will be better as device losses will be less than the second amp also the supply voltage regulation at lesser current will be better .
1, a class AB Nchannel amp with high supply voltage and low current ,
2, another amp with two class AB Nchannel amps in bridge mode having the same number of output devices as the above amp but here in two amps and having the same power output rating as the first amp , in this case the supply voltage will be lower with high current .
I feel that the first one will be better as device losses will be less than the second amp also the supply voltage regulation at lesser current will be better .
lars, I have built the low level boards with almost all using metal film resistors. the store didn't have 10R metal films so I used carbons instead.
although there are some changes to the values. the store didn't have 499R metals so I went with 510R, there is also no 5K so I used 5.1k, also, no 10k so I used 11k. in the filter caps, I couldn't find 47uF 160V so I went with 10uF 160V (I could change that when I do find some) transistors are MPSA93, MPSA43, 2N5551, 2N5401, TIP32/31, I plan to use IRF9640 for the P channel fet.
everything else is the same as schematic.
will those affect performance?
although there are some changes to the values. the store didn't have 499R metals so I went with 510R, there is also no 5K so I used 5.1k, also, no 10k so I used 11k. in the filter caps, I couldn't find 47uF 160V so I went with 10uF 160V (I could change that when I do find some) transistors are MPSA93, MPSA43, 2N5551, 2N5401, TIP32/31, I plan to use IRF9640 for the P channel fet.
everything else is the same as schematic.
will those affect performance?
Those resistors values are rather uncritical, as longs as you have the same value in all positions, those 499R. The 5 k is very uncritical since a pot is in series.
...and the answer to your last question is
...and the answer to your last question is
I did use 11k in all 10k positions, and 510R in all 499R positions. 🙂
just making sure, I might have messed with some critical parts of the circuit but no, I didn't. 😀
just making sure, I might have messed with some critical parts of the circuit but no, I didn't. 😀
also, I forgot to ask. is the BC517 Vbe multiplier required to be thermally connected to the main heatsink? my board layout allows it but I'm not sure if it is necessary.......
I'm gonna run the amps with fan cooling.
I also replaced the 47uF feedback caps to ground with 100uF for lower bass response (completely pointless but it's just me. ) and the 22uF into 33uF nichicons (I already have them, the blue 105°C ones, I also have 22uF but these are cheapies)
I'm gonna run the amps with fan cooling.
I also replaced the 47uF feedback caps to ground with 100uF for lower bass response (completely pointless but it's just me. ) and the 22uF into 33uF nichicons (I already have them, the blue 105°C ones, I also have 22uF but these are cheapies)
ok, I'm beginning to think I'm the only one in this thread and no one's reading.
anyway, I have built a pair of Zeta's which run off +/-50V rails. I set them up on the bench bare with only wires soldered etc. I haven't done much listening as it has too much gain and amplifies the noise of my CD player.
I have a few probs though, first, DC offset. I don't know what's causing it as I have ~800mV on one channel and ~380mV on the other. the trimmer that is supposed to adjust DC offset doesn't seem to have any effect.
also, I'm wondering, how do you set the bias? I was only able to measure voltage across the +rail mosfet source resistors (meaning there is current). by the time I was able to measure some voltage across the -rail mosfets, the +rail mosfets start heating up.
everything else is as per schematic. except that I used BC546 for BC517, 2N5551 for 2N5550, MPSA43 for MPSA42, MPSA93 for MPSA92, IRF9610 for IRFP9240. I shorted out the 5k for bias as I wasn't getting any bias even with the trimmer at min resistance.
anyone know what I should know?
anyway, I have built a pair of Zeta's which run off +/-50V rails. I set them up on the bench bare with only wires soldered etc. I haven't done much listening as it has too much gain and amplifies the noise of my CD player.
I have a few probs though, first, DC offset. I don't know what's causing it as I have ~800mV on one channel and ~380mV on the other. the trimmer that is supposed to adjust DC offset doesn't seem to have any effect.
also, I'm wondering, how do you set the bias? I was only able to measure voltage across the +rail mosfet source resistors (meaning there is current). by the time I was able to measure some voltage across the -rail mosfets, the +rail mosfets start heating up.
everything else is as per schematic. except that I used BC546 for BC517, 2N5551 for 2N5550, MPSA43 for MPSA42, MPSA93 for MPSA92, IRF9610 for IRFP9240. I shorted out the 5k for bias as I wasn't getting any bias even with the trimmer at min resistance.
anyone know what I should know?
N Channel amp!!!!
Hi DJ Quan,
I think you need to replace the BC517 as this is a darlington pair used for the high Amplification,
Where Bc 546 is not a darlington ,Cant be replaced .
MPSA 43/93 should not be a problem, but if you are using higher power try using IRF 9640.
see that you have right zobel network, In the Actual ckt of zeta the C10 +ve is connected wrongly Bet'n R54&C11.
Thanks
Arasuk.
Hi DJ Quan,
I think you need to replace the BC517 as this is a darlington pair used for the high Amplification,
Where Bc 546 is not a darlington ,Cant be replaced .
MPSA 43/93 should not be a problem, but if you are using higher power try using IRF 9640.
see that you have right zobel network, In the Actual ckt of zeta the C10 +ve is connected wrongly Bet'n R54&C11.
Thanks
Arasuk.
I figured that BC517 is just for biasing so it wouldn't cause any probs. bias remains stable at any heatsink temp that I have tested although I would still replace it if I could find any.
I could not find BC517 and IRF9640 so I settled on the mentioned devices. the two versions I'll be making will be this one for +/-50V and another for +/-70V. I wouldn't go any higher.
in the schem that I have, the zobel is correctly connected. I wired mine up with C11 connected to power gnd, then the 2.2R to speaker out.
surprisingly, since I ran out of available output devices, I used a pair of IRFP254 for -rail and IRFP460 for the +rail. and the this module has lower DC offset than the one with all IRFP460.
I'm still wondering why there's DC offset. the NFB should eliminate this to a low level.
I could not find BC517 and IRF9640 so I settled on the mentioned devices. the two versions I'll be making will be this one for +/-50V and another for +/-70V. I wouldn't go any higher.
in the schem that I have, the zobel is correctly connected. I wired mine up with C11 connected to power gnd, then the 2.2R to speaker out.
surprisingly, since I ran out of available output devices, I used a pair of IRFP254 for -rail and IRFP460 for the +rail. and the this module has lower DC offset than the one with all IRFP460.
I'm still wondering why there's DC offset. the NFB should eliminate this to a low level.
Hi djQUAN !
Do you remeber BC517 has Very high DC current gain (min. 30000).
and the BC546 only between ß:200-450.
Data BC546: DC current gain IC = 2 mA; VCE = 5 V:
BC546 MIN. TYP. MAX.
---------------------------
BC546A 110 180 220
BC546B 200 290 450
You need mutch more current for only one BC546,
better use two BC546 as discret darlington Transistor.
See:
- BC517: http://www.semiconductors.philips.com/acrobat_download/datasheets/BC517_5.pdf
- BC546: http://www.semiconductors.philips.com/acrobat_download/datasheets/BC546_547_3.pdf
Do you match any T1|T4 and T2|T3 ?
I also test the T9+T10, T12+T13 for the 'same' DC current gain
I list many comments from Lars C. on my website too: http://bauteile-fuer-die-elektronik.de/projekte/ZETA/showimg.php?file=/00-Schaltplan/kommentare.php
best wishes
- uwe
-= http://bauteile-fuer-die-elektronik.de/projekte/ZETA/ =-
djQUAN said:
Do you remeber BC517 has Very high DC current gain (min. 30000).
and the BC546 only between ß:200-450.
Data BC546: DC current gain IC = 2 mA; VCE = 5 V:
BC546 MIN. TYP. MAX.
---------------------------
BC546A 110 180 220
BC546B 200 290 450
You need mutch more current for only one BC546,
better use two BC546 as discret darlington Transistor.
See:
- BC517: http://www.semiconductors.philips.com/acrobat_download/datasheets/BC517_5.pdf
- BC546: http://www.semiconductors.philips.com/acrobat_download/datasheets/BC546_547_3.pdf
Do you match any T1|T4 and T2|T3 ?
I also test the T9+T10, T12+T13 for the 'same' DC current gain
I list many comments from Lars C. on my website too: http://bauteile-fuer-die-elektronik.de/projekte/ZETA/showimg.php?file=/00-Schaltplan/kommentare.php
best wishes
- uwe
-= http://bauteile-fuer-die-elektronik.de/projekte/ZETA/ =-
djQUAN
arasuk is right BC517 is a darlington transistor , I think BC 547 should work however if it does not make a darlington using two BC 547s
Disconnect the output devices and measure voltage across C8 , does it vary from 5v to 7v by the preset R22
Regarding the offset voltage use matched T1 T4 &T2 T3. Check offset voltage at collectors of TR10 &TR11 . I assume you would have checked all the connections .
arasuk is right BC517 is a darlington transistor , I think BC 547 should work however if it does not make a darlington using two BC 547s
Disconnect the output devices and measure voltage across C8 , does it vary from 5v to 7v by the preset R22
Regarding the offset voltage use matched T1 T4 &T2 T3. Check offset voltage at collectors of TR10 &TR11 . I assume you would have checked all the connections .
I did not match the transistors as I don't have the budget to buy enough (and I don't have the transistor gain tester at hand, gotta get one soon!) I got the BC546B
what about 2N5306? (I think I got that right?) that's the only darlington transistor in a to-92 package that the local store sells.
I'll try to match the transistor gains as best as I can and see what happens.
without changing anything, I'll try to see if the voltage changes at C8 this weekend.
if everything else fails, I plan to add a DC servo or something......
what about 2N5306? (I think I got that right?) that's the only darlington transistor in a to-92 package that the local store sells.
I'll try to match the transistor gains as best as I can and see what happens.
without changing anything, I'll try to see if the voltage changes at C8 this weekend.
if everything else fails, I plan to add a DC servo or something......
I just had a look at the schem, voltage across C8 changes as I can adjust bias. with a few minor mods to be able to use BC546 works.
the only prob is the output DC offset.
also, I tried measuring the dc across the inverting and non inverting input of the amp. the one with 700+mV offset has around 50mV across + and - inputs, the one with 300+mV offset has around 5mV across + and - inputs.
the only prob is the output DC offset.
also, I tried measuring the dc across the inverting and non inverting input of the amp. the one with 700+mV offset has around 50mV across + and - inputs, the one with 300+mV offset has around 5mV across + and - inputs.
Hi djQUAN !
I use this "Full Featured Transistor Tester" ESP Project 31.
it's nice and easy to construct !
In my version on a Testboard,
i can use Ib: 1uA, 10uA and 100uA and
Ic: 1mA, 10mA, 100mA.
Ib: base current
Ic: collector current
all the best from germany
-uwe
djQUAN said:
I use this "Full Featured Transistor Tester" ESP Project 31.
An externally hosted image should be here but it was not working when we last tested it.
it's nice and easy to construct !
In my version on a Testboard,
i can use Ib: 1uA, 10uA and 100uA and
Ic: 1mA, 10mA, 100mA.
Ib: base current
Ic: collector current
all the best from germany
-uwe
Attachments
Hi Folks!
So i'm back from the exposition in a rather beautiful Bucharest, Romania, where i have spent the last 10 days. 🙂
QUAN: You need to use a darlington, or the voltages will not add up to the required level to turn on both MOSFET's With a normal transistor like BC546 the BIAS voltage is too low to turn on both the positive and negative MOSFET.
I am sure your DC offset comes from the negative rail MOSFET's are simply all off. The 9610 has a higher on threshold for the required current, compared to 9640, so this will make the problem of using BC546 even worse.
The reason why i am using a darlington in the first place is that a normal transistor will loose a part of it's Vbe multiplication due to current flowing in the base. This means it will not compensate Rcb/Rbe but only a percentage of that. So to have a reliable Vbe multiplier, you always have to use a darlington.
Try making the discrete darlington using two BC546, it will work fine.
All the best from
Lars 🙂
So i'm back from the exposition in a rather beautiful Bucharest, Romania, where i have spent the last 10 days. 🙂
QUAN: You need to use a darlington, or the voltages will not add up to the required level to turn on both MOSFET's With a normal transistor like BC546 the BIAS voltage is too low to turn on both the positive and negative MOSFET.
I am sure your DC offset comes from the negative rail MOSFET's are simply all off. The 9610 has a higher on threshold for the required current, compared to 9640, so this will make the problem of using BC546 even worse.
The reason why i am using a darlington in the first place is that a normal transistor will loose a part of it's Vbe multiplication due to current flowing in the base. This means it will not compensate Rcb/Rbe but only a percentage of that. So to have a reliable Vbe multiplier, you always have to use a darlington.
Try making the discrete darlington using two BC546, it will work fine.
All the best from
Lars 🙂