FORR
>Beware that the two pole compensation can lead to instabilities............
MIKEKS
>Only if you don't know how to apply it....
I would be glad to know. I exactly used the values as indicated by Self and my boards come from Signal Transfer, which means they are of high standard.
~~~~~~ Forr
§§§
>Beware that the two pole compensation can lead to instabilities............
MIKEKS
>Only if you don't know how to apply it....
I would be glad to know. I exactly used the values as indicated by Self and my boards come from Signal Transfer, which means they are of high standard.
~~~~~~ Forr
§§§
mikeks-
Only if you don't know how to apply it....
hey mike, how do you mean, i dont know
specifics required please, between you and me ive been busy
at my lab applying it and realised that it works however i applied
it,hey mike ive come up with a better solutioin- different compensation!
but i would be most interested to see how you yourself would apply it
so if it is not of inconvinience of eny sort i would welcome any
specifics youd wish to share
cheers
Only if you don't know how to apply it....
hey mike, how do you mean, i dont know
specifics required please, between you and me ive been busy
at my lab applying it and realised that it works however i applied
it,hey mike ive come up with a better solutioin- different compensation!
but i would be most interested to see how you yourself would apply it
so if it is not of inconvinience of eny sort i would welcome any
specifics youd wish to share
cheers
thanks guys for your valuable suggestion. Just finish one channel and tested with +/-24. Both input and output goes to +24V. Can it work for +/-24V?
Andrew,
Can I increase the miller caps to slow down the amp instead of RF filter as I dont like caps along the signal path.
MikeB,
I've replace all small signal to MPSA42/92 but mantain Q5/Q6 as BC560C due to lower voltage there.
Jaycee,
Do you forsee any problem with MJE15030 as VAS? i use it for easy mounting compared to to92 package and also because I have plenty of it. Can you tell me what is the advantage of using BD139 compared to MJE15030?
Mod_evil,
Can't manage to post eagle file here. Can you send me your e-mail address?
thanks.
Andrew,
Can I increase the miller caps to slow down the amp instead of RF filter as I dont like caps along the signal path.
MikeB,
I've replace all small signal to MPSA42/92 but mantain Q5/Q6 as BC560C due to lower voltage there.
Jaycee,
Do you forsee any problem with MJE15030 as VAS? i use it for easy mounting compared to to92 package and also because I have plenty of it. Can you tell me what is the advantage of using BD139 compared to MJE15030?
Mod_evil,
Can't manage to post eagle file here. Can you send me your e-mail address?
thanks.
Ipanema said:
Yes it should work OK.
Ipanema said:
You do realise this is stupid logic. Whether something is visually in the signal path or not, if the end result is the same (i.e. bandwidth limitation) then the sound will be the same. However, increasing the Miller cap is a bad idea because it loads the VAS and input stages.
Ipanema said:
MJE will be fine. Generally using lower power devices yeilds faster performance, but the MJE device is quite modern despite being more powerful, so should be fine.
Hi,
I would not recommend increasing the Miller comp cap.
The input RF filter only rolls off (-3db) from about 200kHz and will be about 1db down at 100kHz and its effect at 20kHz is almost unmeasurable both in response and phase. Since the cap is shunt to earth and is so far out of the audio band it effectively is not in the signal path. The Miller comp cap is very much in the signal path since its effect will be heard/seen from below 100Hz upto MHz
MJL15030 as a VAS will probably not work. It has a Cob=45pF at 30V and will swing from 30pF to 140pF when under load. Not at all good for audio. With this you could probably do without a Miller comp cap since adding one will overload the LTP on higher frequencies. The slew rate and distortion will be pretty poor.
You should be looking for a low Cob transistor with a decent Ft for a VAS. Something like 2sa1478 or 2sa1360 if you need suggestions. Regarding the BD139, I cannot find any decently detailed data sheet and so cannot advise but I can surmise. Since I have seen one uncorroborated Ft=190mHz and it is quoted as 1.5A continuous I would expect the Cob to be similar to 2sb649 at about 6pF to 50pF over it's working range. Still too high.
I would not recommend increasing the Miller comp cap.
The input RF filter only rolls off (-3db) from about 200kHz and will be about 1db down at 100kHz and its effect at 20kHz is almost unmeasurable both in response and phase. Since the cap is shunt to earth and is so far out of the audio band it effectively is not in the signal path. The Miller comp cap is very much in the signal path since its effect will be heard/seen from below 100Hz upto MHz
MJL15030 as a VAS will probably not work. It has a Cob=45pF at 30V and will swing from 30pF to 140pF when under load. Not at all good for audio. With this you could probably do without a Miller comp cap since adding one will overload the LTP on higher frequencies. The slew rate and distortion will be pretty poor.
You should be looking for a low Cob transistor with a decent Ft for a VAS. Something like 2sa1478 or 2sa1360 if you need suggestions. Regarding the BD139, I cannot find any decently detailed data sheet and so cannot advise but I can surmise. Since I have seen one uncorroborated Ft=190mHz and it is quoted as 1.5A continuous I would expect the Cob to be similar to 2sb649 at about 6pF to 50pF over it's working range. Still too high.
Ipanema said:
It's not really about ft, for voltageamplification the outputcapacitance
is dominating. A high outputcapacitance slows down the vas and a
high inputcapacitance puts a heavy load on the previous stage for
higher frequencies.
I think, the mje15030/31 is best used for its original purpose, as
driver in the outputstage...
About the mpsa42/92, now you are save with voltageratings, but you
have to check if they heat up. bd139/140 should be a good choice
as jaycee recommended.
I overlooked your question about the DC-offset...
The current through the bases of the diffamp are quite "high", given
a hfe of ~300, the 3ma create a basecurrent of 10ua, through the
22k resistors this creates a voltage of ~220mv. Luckily, this voltage
is created on both sides of the ltp, as both see a 22k resistor.
But, if you now skip the inputcap, these 220mv on the left side can
be shorted down to 0mv, creating a dc-offset of 0.2v on the output,
depending on the device attached to the input. This can't be good.
Mike
Hi,
my belief, but I have no technical justification, is that the VAS should be faster than the output stage. I further believe that the difference between should be at least one octave.
For 30MHz outputs I would choose a VAS higher than 60MHz and preferably two octaves above.
I have seen quite a few designers choosing VAS at Ft= 200MHz upto 500MHz. Could this be to achieve that very low Cob? Which is almost inherent in the very high Ft transistors.
If you do apply Miller comp cap then high Ft ensures high feedback around that one stage. Similarly capacitor feedback to the LTP -ve will also be high. These topologies rely on high feedback for low distortion.
Q1. What difference in sound quality or measurements will a 30MHz or 100Mhz VAS make?
Q2. What difference will modulated Cob=5pF or 60pF at the VAS stage make to the sound quality?
my belief, but I have no technical justification, is that the VAS should be faster than the output stage. I further believe that the difference between should be at least one octave.
For 30MHz outputs I would choose a VAS higher than 60MHz and preferably two octaves above.
I have seen quite a few designers choosing VAS at Ft= 200MHz upto 500MHz. Could this be to achieve that very low Cob? Which is almost inherent in the very high Ft transistors.
If you do apply Miller comp cap then high Ft ensures high feedback around that one stage. Similarly capacitor feedback to the LTP -ve will also be high. These topologies rely on high feedback for low distortion.
Q1. What difference in sound quality or measurements will a 30MHz or 100Mhz VAS make?
Q2. What difference will modulated Cob=5pF or 60pF at the VAS stage make to the sound quality?
A low Cob is desirable in the VAS because it is non-linear and a primary factor in phase intermodulation distortion. The lower this value, the higher the percentage of total capacitance that an added parallel compensation cap is to the total capacitance. This added cap is much more linear than Cob, and thus reduces the amount off phase intermodulation distortion caused by this capacitance.
Hi Pooge,
"A low Cob is desirable in the VAS because it is non-linear and a primary factor in phase intermodulation distortion. The lower this value, the higher the percentage of total capacitance that an added parallel compensation cap is to the total capacitance. This added cap is much more linear than Cob, and thus reduces the amount off phase intermodulation distortion caused by this capacitance."
Best to eliminate the Vas and it's problems.
Cheers,
Greg
"A low Cob is desirable in the VAS because it is non-linear and a primary factor in phase intermodulation distortion. The lower this value, the higher the percentage of total capacitance that an added parallel compensation cap is to the total capacitance. This added cap is much more linear than Cob, and thus reduces the amount off phase intermodulation distortion caused by this capacitance."
Best to eliminate the Vas and it's problems.
Cheers,
Greg
amplifierguru said:Hi Pooge,
"A low Cob is desirable in the VAS because it is non-linear and a primary factor in phase intermodulation distortion. The lower this value, the higher the percentage of total capacitance that an added parallel compensation cap is to the total capacitance. This added cap is much more linear than Cob, and thus reduces the amount off phase intermodulation distortion caused by this capacitance."
Best to eliminate the Vas and it's problems.
Cheers,
Greg
Yes, it's a nuisance of a problem but it's possible to work with it. A good device is the Philips BF469 / 470. Don't forget too that the VAS current source device needs also to be low Cob. The situation is not necessarily quite as bad as some make out because the Cob of the VAS and current source are always in opposition, so if their curves are near identical then capacitance modulation is nearly eliminated.
I managed to achive vastly superior results by using the lowest Cob / Cib devices possible. Use of MPSA18 in the CM is very recommended.
Unfortunately, from my experience anyway, achieving great results with a single ended IS/VAS design requires restricting it to relatively low power. Great for a preamp or tweeter amp though.
Hi everyone,
My one channel blameless amp has finally managed to open her mouth to sing last night. Thanks to all of you who have given me advice. The amp is dead quite, only week hissing when I place my ear to the speaker cone. I replace C13 with 300pF and left C3 out as I can't find 16pF film type cap. The amp is stable, no sign of oscillation. So do I still need to put in C3?
Also made mono comparison with LM3875 GC. Blameless excell with vocal as the female vocal comes out more relax and smoother compared to GC which is thicker but more full bodied. It is even better with Mark Knopfler vocal, everything sounds smooth and relax.
Blameless is weaker in lower end extension. Suprisingly, i'm using 6800uF per rail snuberized power supply for blameless and only 1000uF for GC. No relation between bass and capacitance value?
The top end will be quite dominating sometimes if the music itself has lots of sharp edges, especially with pop song. With simple jazz/country song it can project more sense of detail. increasing C13 to 600pF reduces the top end sparkling but the overall sound become a bit muddy. 1200pF kill the sound. simple tone control.
I'll try to replace MJE15030 with BD139 as suggested and also two pole compensation. Pls suggest any further improvement that I can make. How to implement seperate current source for LTP and VAS? Will this improve things?
Oh yeah, I have a scope and signal generator. What should I measure?
thanks.
My one channel blameless amp has finally managed to open her mouth to sing last night.
Thanks to all of you who have given me advice. The amp is dead quite, only week hissing when I place my ear to the speaker cone. I replace C13 with 300pF and left C3 out as I can't find 16pF film type cap. The amp is stable, no sign of oscillation. So do I still need to put in C3? Also made mono comparison with LM3875 GC. Blameless excell with vocal as the female vocal comes out more relax and smoother compared to GC which is thicker but more full bodied. It is even better with Mark Knopfler vocal, everything sounds smooth and relax.
Blameless is weaker in lower end extension. Suprisingly, i'm using 6800uF per rail snuberized power supply for blameless and only 1000uF for GC. No relation between bass and capacitance value?
The top end will be quite dominating sometimes if the music itself has lots of sharp edges, especially with pop song. With simple jazz/country song it can project more sense of detail. increasing C13 to 600pF reduces the top end sparkling but the overall sound become a bit muddy. 1200pF kill the sound. simple tone control.
I'll try to replace MJE15030 with BD139 as suggested and also two pole compensation. Pls suggest any further improvement that I can make. How to implement seperate current source for LTP and VAS? Will this improve things?
Oh yeah, I have a scope and signal generator. What should I measure?
thanks.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.