Talking about capacitors there is a need to select the right ceramic di-electric absorption for the job -one where this characteristic is linear across the high frequency range.
I did spot a blue monolithic type in one of your pictures it was not a brown ceramic with a black crescent at the top - a cheap but linear in terms of absorption as above.
If you used a monolithic type for C5 the dominant pole for your amplifier this may measure 27pf under static conditions. What is is more important is what is di-electric absorption factor under dynamic conditions be which needs to be as constant as can be made.
If you have replaced a brown C5 (with the black cap) in the misbehaving amplifier with a blue monolithic which could account for your problem. I suggest looking at C5 in the good amplifier to see what type that is.
If C5 is a brown cermic cap of 27pf you could think about replacing with 33pf.
Last edited:
if you have confirmed that you have MHz signals coming out the amp, then increase those 27p caps to whatever small value you have just to confirm the issue
In the last regard it should be remembered that the collector to base junction capacitance is dependent on the voltage applied to the collector.
Moreover the base to collector of the transistor is reverse biased so there exists a diffusion capacity which is directly and inversely proportional to the square root of the voltage applied to the such junction.
Changing C5 to 33pf is a tentative step which may lead nowhere. particularly if the small signal transistors at the input are high voltage types with low junction capacitance remembering this is a high power amplfier. Ceramic brown caps are not available beyond 100pf where I live.
If one is to progress in the manner suggested there are strings of pin holders which could be soldered to a pcb snapping -two of which could allow testing for a range of capacitors.
Moreover the base to collector of the transistor is reverse biased so there exists a diffusion capacity which is directly and inversely proportional to the square root of the voltage applied to the such junction.
Changing C5 to 33pf is a tentative step which may lead nowhere. particularly if the small signal transistors at the input are high voltage types with low junction capacitance remembering this is a high power amplfier. Ceramic brown caps are not available beyond 100pf where I live.
If one is to progress in the manner suggested there are strings of pin holders which could be soldered to a pcb snapping -two of which could allow testing for a range of capacitors.
Last edited:
I should add that C7 should be included in frame with C5 these are BF469 transistors - high voltage 60MHz types. I thought these were no longer current from the original sources but are manufactured by other producers. It would be interesting to see the faces of the the actual devices for some identity clues.
Last edited:
Mjona, Citrix, thanks
The 27pF caps are the tiny brown ceramic types with black identification on both. I have measured them off the board in both amps all very close to 27pF
The only blue cap is a small film in the anti thump circuit
It does look like one of the Buz 905 mosfet is failing in the oscillating amp as various measurements around the amp are different with one of these in place
It is just under 4KHz resistance across the drain and gate pins so I will need to swap this out as the others all measure open cct
I’m not sure but suspect I’ll need to measure some or all of the mosfets to match accordingly ? Perhaps buying a suitable peak electronic gauge
I do have a 22pF cap (Polystyrene) and might have a 39pF somewhere (naim use them) so I can tack them to the back of the 27pF caps to be in parallel
Given one of these mosfets is showing issues, I’d assumed that would be the first thing to sort ?
To jean-Paul’s delight Claude suggested it might !
BTW I use copper bar to eliminate fuse issues when testing sometimes. This amp has two out of spec fuses not 15A and they are different as well so I will get the right spec ones. I’m also thinking about replacing with some auto reset ones as the 500 amp has these and I just wanted to see if there might be a fuse issue.
Adrian
The 27pF caps are the tiny brown ceramic types with black identification on both. I have measured them off the board in both amps all very close to 27pF
The only blue cap is a small film in the anti thump circuit
It does look like one of the Buz 905 mosfet is failing in the oscillating amp as various measurements around the amp are different with one of these in place
It is just under 4KHz resistance across the drain and gate pins so I will need to swap this out as the others all measure open cct
I’m not sure but suspect I’ll need to measure some or all of the mosfets to match accordingly ? Perhaps buying a suitable peak electronic gauge
I do have a 22pF cap (Polystyrene) and might have a 39pF somewhere (naim use them) so I can tack them to the back of the 27pF caps to be in parallel
Given one of these mosfets is showing issues, I’d assumed that would be the first thing to sort ?
To jean-Paul’s delight Claude suggested it might !
BTW I use copper bar to eliminate fuse issues when testing sometimes. This amp has two out of spec fuses not 15A and they are different as well so I will get the right spec ones. I’m also thinking about replacing with some auto reset ones as the 500 amp has these and I just wanted to see if there might be a fuse issue.
Adrian
Hi, I see you mentioned my name but I do not see delight? The logic and the fear of being ridiculed go hand in hand when you write:"BTW I use copper bar to eliminate fuse issues when testing sometimes.". Normally fuses are used specifically when testing to avoid issues 😉 They protect wiring and meltdown.
It was just a hint to prevent you from being ridiculed as technical persons as found here usually are not very enthusiastic seeing dangerous electrical practices they are trained to NOT do. Even in emergency situations such is not usual habit. I once had bolted in a 1250A fuse for a 630A one (instead of the worst: a short) as the device had to function but I took full responsibility and monitored the situation myself and made sure the right part was there the next day.
You have just under 4KHz resistance across the drain and gate pins of the MOSFET which is not OK. If it is 4 kOhm it is also not OK but it is better to measure such parts out of circuit. It would not be the first time in circuit measuring gave erroneous results.
It was just a hint to prevent you from being ridiculed as technical persons as found here usually are not very enthusiastic seeing dangerous electrical practices they are trained to NOT do. Even in emergency situations such is not usual habit. I once had bolted in a 1250A fuse for a 630A one (instead of the worst: a short) as the device had to function but I took full responsibility and monitored the situation myself and made sure the right part was there the next day.
You have just under 4KHz resistance across the drain and gate pins of the MOSFET which is not OK. If it is 4 kOhm it is also not OK but it is better to measure such parts out of circuit. It would not be the first time in circuit measuring gave erroneous results.
Last edited:
Whoops yes my typo 4KOhm I think it makes sense to pull one or two off the board but a fair bit of heatsink dismantling so after Easter probably
Perhaps pulling all 6 of the board to measure all the 905’s on that amp
Not sure if one of these might be worth investing in ?
https://www.peakelec.co.uk/acatalog/dca55-atlas-dca-semiconductor-analyser.html
Perhaps pulling all 6 of the board to measure all the 905’s on that amp
Not sure if one of these might be worth investing in ?
https://www.peakelec.co.uk/acatalog/dca55-atlas-dca-semiconductor-analyser.html
Seriously?
At the risk of losing diplomacy...and some darwin awards candidate....
Time to 'hit the books' me friend, and move on from the drinkie drinks me thinks.
Seriously.
A word about multiple MOSFET I recall reading somewhere by Linsley-Hood is that care is needed in how these are physically arranged. From what I remember about this was that there could be cross-feed between individual devices in doubling up cases. He used Hitachi ones in his designs. You have two amplifiers one being faultless and the other you are less happy with. You could could compare the MOSFET wiring to see if there is any physical wiring feed differences between the two. This could be a reason for your 905 casualty.
Last edited:
I think there is a major point that has escaped notice which is the source follower output stage are somewhat susceptible to oscillation and must be protected from capacitive loads. To mitigate that risk
A slight digresion you might like to check out with other MOSFET projects on this site
While your MOSFET devices have gate stopper resistors that does not isolate the source electrode terminals from capacitative loads. There is nothing to prevent individual devices in this circuit seeing a capacitative load. There should be low value wire-resistors which introduce a small level of inductance in each source chain.
Mjona, thanks
I will hopefully have some time for more reading and investigation today, I found this paper so will have a read here as well https://www.passdiy.com/pdf/mos.pdf
The boards and physical layout is identical apart from the revised anti thump circuit.
I will hopefully have some time for more reading and investigation today, I found this paper so will have a read here as well https://www.passdiy.com/pdf/mos.pdf
The boards and physical layout is identical apart from the revised anti thump circuit.
I did mean wire-wound resistors. With the number of output MOSFETs you have can afford greater resistance which implies more inductance in the wire-wound range. If you are tight for space 2.5 W resistors should suffice. I did note in the Pass link a preference for fewer parts where there is less to go wrong.
The footprint of your output devices is larger than it would be with what Pass describes as Poppa devices. These will be dearer to buy and there are benefits in buying components in bulk since that is cheaper. If you fit wire-wound source resistors you will have to experiment with the values. you will have to experiment. 0.47 Ohm could be a starting point.
The footprint of your output devices is larger than it would be with what Pass describes as Poppa devices. These will be dearer to buy and there are benefits in buying components in bulk since that is cheaper. If you fit wire-wound source resistors you will have to experiment with the values. you will have to experiment. 0.47 Ohm could be a starting point.
Mjona are you suggesting adding some resistors between the case (source) and the tracks they mount too ? That will be tricky as they bolt through to connect to the tracks ?
In your early pages I see some Exicon ECF MOSFETs in metal cans. These are not shown in the recent circuit diagram you posted. The "ECF"amp seems to be a separate animal from the multi BUZ plastic MOSFETs in the circuit diagram that you are having is with.
It should be easier to fit series wire wound source resistors with metal cans with bolt insulation washers and solder tags. Are we looking at two animals in one box as one unit or two.
It should be easier to fit series wire wound source resistors with metal cans with bolt insulation washers and solder tags. Are we looking at two animals in one box as one unit or two.
Last edited:
Hi the other picture was of the BPSP 500 these look to be SML Buz 900, 905 you need a mirror and torch to read them until the they are ff the chassis
I bet it costs more than a fuse. I don't have coper rods laying around and they happen to be the standard fuse size..
Your Exicon ECFN20 N/ECFP20 in TO3 cans are not direct equivalents as replacements for Hitachi 2SK176/SJ56 but according to the sales department information these will suit most circuits. It might be wise not to mix ECF devices with BUZ types.
While an increase in the subwoofer might indicate an improvement in sound there are electro-mechanical forces which might mean more is less. One assumes you have investigated the relevant speaker/new box parameters.
While an increase in the subwoofer might indicate an improvement in sound there are electro-mechanical forces which might mean more is less. One assumes you have investigated the relevant speaker/new box parameters.
Finally starting to make some progress, I have added an old pre amplifier to work with the BK 1000W amps and realised that the BK BPSP 500 is now similar with no active filter (I use this option with the BPSP 500 and with it its silent).
Just measuring noise with no signal, all connected and on. The green is the BK 500 with the frequency contour (roll off) bypassed the black with it in place and the 'quiet' target. note the spec of the 500 says bypassed 3Hz to 1KHz so its rolled off anyway over 1KHz. The brown is the BK 1000W amp with a pre amplifier in series (the 500 has a pre amplifier and vol.)
So the answer is replicating the frequency roll off, I dont need it adjustable like the 500 is (40Hz-120Hz) I set mine at 100Hz
Looking closer at the 20-1KHz region. The pink trace is the cheap DAC with no connections which seems to add a lot of the mains related peaks, which I don't hear. In reality the green trace is as noisy as the orange (BK BPSP 500 with filter bypassed- green) (BK1000W with a pre amplifier in series-Orange), audible noise at the seating position. The black trace is total silence
It's this circuit without the need for adjustability I need to replicate (highlighted in red), any advice here as I haven't built any Opamp based circuits before ? Is this a second order low pass in series with another to make it 4th order ?
Just measuring noise with no signal, all connected and on. The green is the BK 500 with the frequency contour (roll off) bypassed the black with it in place and the 'quiet' target. note the spec of the 500 says bypassed 3Hz to 1KHz so its rolled off anyway over 1KHz. The brown is the BK 1000W amp with a pre amplifier in series (the 500 has a pre amplifier and vol.)
So the answer is replicating the frequency roll off, I dont need it adjustable like the 500 is (40Hz-120Hz) I set mine at 100Hz
Looking closer at the 20-1KHz region. The pink trace is the cheap DAC with no connections which seems to add a lot of the mains related peaks, which I don't hear. In reality the green trace is as noisy as the orange (BK BPSP 500 with filter bypassed- green) (BK1000W with a pre amplifier in series-Orange), audible noise at the seating position. The black trace is total silence
It's this circuit without the need for adjustability I need to replicate (highlighted in red), any advice here as I haven't built any Opamp based circuits before ? Is this a second order low pass in series with another to make it 4th order ?
Wow what a thread....my $0.02 .... make sure signal ground is only connected to chassis ground at one point. T'eed off the common capacitor connection as has been re iterated several times. Connect the AC power input with earth ground connected (not lifted) Then short the input RCA with the typical output impedance from a pre amp driving stage say 1k ohm. If the output is then too noisy for you...trash it, get a new amp. Some amps designed for high power sound reinforcement are just not suitable from home hi-fi. Especially if you are fussy about residual noise level...which I am also, VERY FUSSY.
I think the passive RC input filter is a weak point in the unaltered circuit. To make this more effective one could simply change the component values of this filter. However I don't think this or op.amp filtering would quell frequencies in the radio spectrum. I have tried including an anti parasitic ferrite bead FX1115 and passing the insulated signal wire through the central hole and including this part in series with the input. For more detail on that part you can find info on the internet.