| camt1 |
hello, i have been given an old amplifier made by Burdett and Dales of Sidcup in Kent. It proclaims to be 100w sterio,its old at least 15 years. So far i have replaced some burnt out resistors and the mains wiring and it works ok (driving 2 10"subwoofers no problem) .My question is , is what do you guys think i could do to improve this circiut . The only thing that concerns me so far is the output transistors dont seem to get warm at all but the driver trans (Q4 and 5 ) get very hot to the touch (no heat sinks on (Q4.5.6.7) ,could there be a problem with the bias ?.I wondered if maybe it would be worth replacing the transistors and caps for new and better types. I have included the circiut diagram any ideas and thoughts would be welcome.
thanks camt |
|
|
| Wavebourn |
| Your schematic contains errors, mostly emiotters and collectors are swapped on some transistors. |
|
|
| richie00boy |
Make that most transistors :)
To be honest, messing about modding stuff like this is generally a waste of time. You'd be better off just re-using key parts like the power supply and heatsinks, case, maybe output transistors, to make a better design. Although having said that, the schematic you've drawn (I can see through the wrong component orientations) is not a really bad design by any means. |
|
|
| camt1 |
| sorry corrected drawing my mistakes not used to ms viso |
|
|
| camt1 |
| schematic hopefully right this time |
|
|
| Wavebourn |
| Yes, it is right now, and pretty good classical design for 8-16 Ohm load. |
|
|
| darkfenriz |
This amplifier has no thermal sensing, so works with no questient current in output stage to prevent thermal runaway.
Replace 47R with Vbe multplier circuit mounted on main heatsink and set questient current of output stage to some 15-30mA.
You can also change electrolytic cap in feedback path (22u) to bigger one bypassed with 100n quality cap. |
|
|
| john_ellis |
You could try putting a Vbe multiplier in place of the 47 ohm bias resistor. You don't say if the drivers are on the same heatsink as the output pair, but I assume they aren't.
There has been a lot of discussion about bias on the Class B thread....
2N3772's are a bit old. They may be slow and you could swap them for a modern pair of MJ21194's or if you are willing to hack the board a little more a complementary pair of MJ21193/4's.
TIP3x's don't make good drivers- they're slow too. So if you want a bit of response try swapping these for an MJE15030/31 pair or similar. And then you could change the VAS and current source to faster devices too...
It all depends how much you want to upgrade!
John |
|
|
| camt1 |
Thanks for the replies , only the output tansistors are on heatsinks (large 4" x 4" one for each transistor ). The drivers have no heatsinks and Q4 and 5 run extreamelly hot while although the others even under hard use seem to stay cold.This seems to be different to every other amp i have used . With no quiscent current bias and operating in class B i assume that the crossover distortion will be high. would i be right in thinking of the following values for the VBE multiplier R1 1K , R2 2K variable pot and a BD139 for the transistor mounted on the heatsink any comments would be very helpfull
thanks camt |
|
|
| phase_accurate |
It looks as if this amp behaves like an Edwin amp where only the drivers conduct at low output currents. This correlates well with your observation of drivers being hotter than the output devices themselves.
Regards
Charles |
|
|
| ilimzn |
| quote: | Originally posted by camt1
With no quiscent current bias and operating in class B i assume that the crossover distortion will be high.
|
You assume correctly.
| quote: | | Would i be right in thinking of the following values for the VBE multiplier... |
A bit difficult to say anything about this since you failed to provide a schematic for the Vbe multiplier which would tell us which resistors R1 and R2 are.
In your schematic, the Vbe multiplier needs about 3x Vbe across it, and also an ability to start with lower than that, i.e. zero bias current (provides a safe setting for powering up the amp first time).
Try this: connect 2.2k across BD139 C and B. Then connect a 2k pot's outer contacts from B and E. Connect a 1.8k between the BD139 E and wiper of the pot. If you need a higher maximum idle current, you can reduce the 1.8k to a 1.5k.
A further possible improvement would be the inclusion of a Baxandal diode in the emitter circuit of Q7 (search this forum for details). This will require the Vbe multiplier to give approx. 4x Vbe maximum across it, so the resistor between wiper and E of BD139 should be reduced to 820 ohms to 1k.
| quote: | Originally posted by phase_accurate
It looks as if this amp behaves like an Edwin amp where only the drivers conduct at low output currents. This correlates well with your observation of drivers being hotter than the output devices themselves.
|
Actually, no - the Edwin had a relatively high bias class A driver stage that was also conected via a resistor to the output. Similar to forum member Wavebourn's Swinik. |
|
|
| darkfenriz |
| Resistor at Q2 collector is a bit too high. |
|
|
| Tube_Dude |
After the implementation of the Vbe multiplier, as Ilimzn told you , is a good idea to diminish the value of the resistors at the output from 1 Ohm to something like 0,2 , 0,33 Ohms.
Also, put a bypass cap across the Vbe multiplier circuit ... |
|
|
| camt1 |
hi thanks for the replies . I have modifed and installed the VBE circiut , i have set the voltage drop across the output transistor resistors at 10mv each . I removed one leg of one resistor and found that they are not 1R but labelled as W21 R10 this if memory serves me right means they are 0.1 reistance , which means that quiscent current is 100ma if im right, is this too high ? . Output transistors warm but not hot , have tested underload with no problems of over heating and appears to be stable. I have one further question , one channel is a lot quiter than the other with DC on the output of 350mvcompared to the other channel which is less than 10mv . I have checked over the boards and cannot find any differnces with voltages or components from one side to the other , does anybody have an idea what i should be looking for to make the gain the same as the other board. What values should be the feedback capacitor (at prensent its 22uf ) and the cap accross Q5 base collector (no markings to be seen ) thanks for any help
camt |
|
|
| richie00boy |
| The feedback cap is a bit too small at the moment, but increasing it will make DC offset worse. I do believe there is a fault with the channel showing more offset. >100mV is too much. Has the amp been repaired at all or subjected to any 'bad' conditions such as shorted speaker wires or pops or thumps e.g. by plugging/unplugging live sources? |
|
|
| Tube_Dude |
| quote: | Originally posted by richie00boy
The feedback cap is a bit too small at the moment, but increasing it will make DC offset worse. |
The value of the feedback cap has noting to do with the offset voltage.
But the cap can be leaking and if so ,it can be the culprit of the offset issue...
Inspect also the input capacitor for leaking....... |
|
|
| camt1 |
Yes amp has been repaired at some time all trans except Q1.2.3 have been replaced , the only change being TIP32C used in place of orignal TIP30C (different hfe seems to be the only difference). Replaced i/p cap and feedback cap no change to DC offset or to lack off gain. Would BC448 be a suitable replacement for Q1.2.3 BC477 as there are the only orignal trans left and mayybe suspect.
thanks camt |
|
|
| camt1 |
thanks for the help replaced feedback cap with a 47uf from 22uf this seems to have fixed the gain issue
camt |
|
|
| Tube_Dude |
| quote: | Originally posted by camt1
Would BC448 be a suitable replacement for Q1.2.3 BC477 as there are the only orignal trans left and mayybe suspect.
thanks camt |
The voltage ratings (80volts) are a little in the low side for your's +- 50 volts rails.
It can work, but with a 100 volts transistor you will have more piece of mind... ;)
First try to swap the transistors from one channel to the other and check if the problem follow the change...
Here is the ratings of the BC448
http://www.datasheets.org.uk/specsheet.php?part=BC448
| quote: | | replaced feedback cap with a 47uf from 22uf this seems to have fixed the gain issue |
The value of the capacitor ,as I have already told you ,don't have nothing with the gain in the audio band( only in the very low frequencies ~10Hz).
If the gain have changed , proved that the original capacitor was bad... |
|
|
| camt1 |
sorry should have explained amp is only being used to drive sub woofers from 90Hz down over , it is being fed from linkwitz rielly crossover and linkwitz transform circiut so it was only at low frequnecies that the lack fo gain was noticable .
camt |
|
|
| Tube_Dude |
| quote: | Originally posted by camt1
sorry should have explained amp is only being used to drive sub woofers from 90Hz down over , it is being fed from linkwitz rielly crossover and linkwitz transform circiut so it was only at low frequnecies that the lack fo gain was noticable .
camt |
Ok,;)
In that case , I will try 100 or 200 uF... :cool: |
|
|
| camt1 |
thanks for the help every body one old amp up and running and sounding good ,thanks tube dude fitted 100uf cap , now the floors shaking
camt |
|
|
|
