Hi Guys
The base-stops on the control BJTs in the current sources are useful as I described earlier. They also protect the base in the event of certain circuit failures. Since others suggest having provision for base-stops in other positions, it does not hurt to have the same provision here.
Regarding the modified BOB400 of post-40: The output stage looks similar to the Crown DC150, They biased the output devices 'off' and used identical devices for the drivers. Once the load power got to a current high enough to turn on the outputs, they contributed to driving the load. Self had many negative things to say about this, where Rod Elliot made a 500W sub amp using the concept.
I made an EF4 in a similar vein, with 1302/3281 for drivers and outputs. The drivers idle at 12W each on 100V rails, with the outputs at 6W each. THD20 is an order of magnitude lower than you posted, at 90Vpk into loads down to 2R. Lots of output transistors...
Crossing the predriver emitter resistor references is not necessarily needed. Did you find it made a difference?
Have fun
The base-stops on the control BJTs in the current sources are useful as I described earlier. They also protect the base in the event of certain circuit failures. Since others suggest having provision for base-stops in other positions, it does not hurt to have the same provision here.
Regarding the modified BOB400 of post-40: The output stage looks similar to the Crown DC150, They biased the output devices 'off' and used identical devices for the drivers. Once the load power got to a current high enough to turn on the outputs, they contributed to driving the load. Self had many negative things to say about this, where Rod Elliot made a 500W sub amp using the concept.
I made an EF4 in a similar vein, with 1302/3281 for drivers and outputs. The drivers idle at 12W each on 100V rails, with the outputs at 6W each. THD20 is an order of magnitude lower than you posted, at 90Vpk into loads down to 2R. Lots of output transistors...
Crossing the predriver emitter resistor references is not necessarily needed. Did you find it made a difference?
Have fun
Base stoppers on the current source BJTs add to the stability of the current source by increasing the PM of the local control loop. They are optional in my opinion. Personally, haven't used them to date. Suspect there may be a penalty in another spec of the current source as there is no "free lunch" with these things.
Paul
Paul
Are you saying the base stoppers are on the wrong two transistors?
I have seen a few Members saying that base stoppers are needed on BJT CCS. But I can never remember which transistor (or both) needs such.
And I can never find the information with a search.
If two CCSs of that type share a common reference, they can go out of whack if one of them saturates (runs out of voltage compliance). Putting a base resistor in there prevents problems. It's not really a base *stopper* per se, because it has nothing to do with local loop stability at high frequency. It's more of a "DC stability" thing.
Base stoppers on the other two would help loop stability, if there were an issue. With that type of CCS, I've never seen local oscillation issues. That doesn't mean it's impossible.
Base stoppers on the other two would help loop stability, if there were an issue. With that type of CCS, I've never seen local oscillation issues. That doesn't mean it's impossible.
He has changed from the common supply after Struth suggested the change.
They now share only up to the first 10k + cap. Then isolated with two 15k, one to each CCS. post 34 shows.
They now share only up to the first 10k + cap. Then isolated with two 15k, one to each CCS. post 34 shows.
Last edited:
Hi guys,
Referring to schematics in post #38...
I noticed something. If I remove Q12, the "VAS protection" transistor, I got positive rail on output. If I put 2,2k resistor in Q11 collector the problem is solved, but I'm wondering if there is a better solution...
Referring to schematics in post #38...
I noticed something. If I remove Q12, the "VAS protection" transistor, I got positive rail on output. If I put 2,2k resistor in Q11 collector the problem is solved, but I'm wondering if there is a better solution...
Last edited:
Resistor in the collector of the beta enhancer is a good idea anyway. The only time you can really get away without it is if you hook up the VAS as a darlington (connect the collectors of the VAS proper and the beta enhancer Q11).
Looks good to me.
I think most of the bases have been covered.
I usually use transistors at half their spec'ed power rating.
A little bit of heat is ok. I start to get a bit worried if they are more than 50-60 degree's C.
I have a cheap IR thermometer which has been very useful.
I think most of the bases have been covered.
I usually use transistors at half their spec'ed power rating.
A little bit of heat is ok. I start to get a bit worried if they are more than 50-60 degree's C.
I have a cheap IR thermometer which has been very useful.
I want to put 2 x 0,68R/2W in emitter of the output devices and I found to buy only metal oxide film resistors. I searched the net about this resistor type and I found this. Here it says that: "The metal oxide film resistors have poor properties for low values and tolerances". Mine are 0,68R so this is my case. But I wonder what kind of properties is referring to ? What is so worst about these low value metal oxide resistors ? Are they worst than clssical wirewound resistors ?
Last edited:
I think Ayrton Perry wirewounds are supposed to be the better choice for emittor resistors.
http://www.vishay.com/docs/31801/mra.pdf
http://www.vishay.com/docs/31801/mra.pdf
metal oxide generally have a higher tempco and a very slightly higher noise.
But in their favour they have a higher overload tolerance compared to thin film/metal film.
Accuracy is easily solved by matching from the batch you buy.
You will need at least 16 for a pair of amplifiers. That leaves a lot of leaway to select out good matches for paralleled pairs/triples.
A Member in another Thread dared to ask if I was against metal oxide in audio circuits. Here's am excellent example where oxide will likely perform better.
But in their favour they have a higher overload tolerance compared to thin film/metal film.
Accuracy is easily solved by matching from the batch you buy.
You will need at least 16 for a pair of amplifiers. That leaves a lot of leaway to select out good matches for paralleled pairs/triples.
A Member in another Thread dared to ask if I was against metal oxide in audio circuits. Here's am excellent example where oxide will likely perform better.
So you say it's ok if I use metal oxide in emitter resistor of power stage ?
In normal operation they don't heat up too much. With sinusoidal signal they dissipate 0,65W and their power rating is 4W.
In normal operation they don't heat up too much. With sinusoidal signal they dissipate 0,65W and their power rating is 4W.
In normal operation they don't heat up to any significant degree.
A 100W into 8ohms 2pair output stage, operating at an average level of -20dB passes 350mAac to the load. For half the time either the upper, or lower, half pass this current, i.e. duty = 50%.
While passing this current they share it.
So we have an effective 350/4 = 88.4mArms through each emitter resistor.
The dissipation is 0.0884² * 0r34 = 2.6mW shared between two 0r68.
A 2W 0r68 remitter resistor is operating at 1.3% of it's rating. You won't feel any heat.
Transients by nature have a short duratiion. Therefore no signifiacnt heating.
Do the numbers for your power output into your load and try for -20db and also -10dB of average levels.
I just did 800W into 4r0, 3pair output stage, operating at -10dB and the 0r68 2W emitter resistor is dissipating 1.2% of maximum rating. Check this number, I haven't.
A 100W into 8ohms 2pair output stage, operating at an average level of -20dB passes 350mAac to the load. For half the time either the upper, or lower, half pass this current, i.e. duty = 50%.
While passing this current they share it.
So we have an effective 350/4 = 88.4mArms through each emitter resistor.
The dissipation is 0.0884² * 0r34 = 2.6mW shared between two 0r68.
A 2W 0r68 remitter resistor is operating at 1.3% of it's rating. You won't feel any heat.
Transients by nature have a short duratiion. Therefore no signifiacnt heating.
Do the numbers for your power output into your load and try for -20db and also -10dB of average levels.
I just did 800W into 4r0, 3pair output stage, operating at -10dB and the 0r68 2W emitter resistor is dissipating 1.2% of maximum rating. Check this number, I haven't.
Last edited:
I start working on the PCB.
Do you think the power stage it's ok like this ?
And from where should I take the NFB signal, from the close left or the far right near the output coil ? I know it's better to take it from the output but the NFB trace will be pretty long (about 10 cm). This will not insert some inductance ?
Do you think the power stage it's ok like this ?
And from where should I take the NFB signal, from the close left or the far right near the output coil ? I know it's better to take it from the output but the NFB trace will be pretty long (about 10 cm). This will not insert some inductance ?
An externally hosted image should be here but it was not working when we last tested it.
Unless you are planning to use massive heatsinks, your center output devices are going to run warmer than the outer ones, and may start to current hog. They should be spaced farther apart.
Take NFB from right beside the output inductor. NFB needs to see exactly the same as what's going out to the speaker to work best.
Take NFB from right beside the output inductor. NFB needs to see exactly the same as what's going out to the speaker to work best.
A trace past all those high current rails isn't ideal. In the Slewmaster designs, NFB was taken to the input through an external wire, away from any high current. Not ideal either, but it worked well. Another option would be to move the output inductor closer to the center.
I have another question. This is the schematics.
I don't know which type of diode is better to use in the supply rejection filter D2+R53+C19: normal rectifier, ultrafast, schottky.... Until now have used ultrafast but maybe a slow diode will reject better high frequencies. What do you think ?
I don't know which type of diode is better to use in the supply rejection filter D2+R53+C19: normal rectifier, ultrafast, schottky.... Until now have used ultrafast but maybe a slow diode will reject better high frequencies. What do you think ?
- Status
- Not open for further replies.