Power amp under development

Jump to Latest
#1,343
Re[15]: Mini - Quasi

quasi said:


Re: Filtering capacitors. It's all about space available on the board and diminishing returns. If you can get a larger value capacitor to fit then by all means put it in. Much larger value capacitors can be found for rails under 50v.
Click to expand...

Let's just say I one of the modifications I have done to the base board has been "creative" in this regard 😉

I am not saying anything specific on what modifications I have created thus far in theory with the PCBs in part as if people want to change something they can change what they wish which may not be same as what I wish or compatiable with. So of my changes are driven by the part size I source that is not compatiable with the board. The options for others include using a graphics editing program, as I have done, or they can buy the PCB CAD program you use and then make the changes again as some have chosen to do as well. I do not want to encourge people to make too many requests for board changes for you to make for them. That is why I have not said directly too much about the changes I have effected or yet to effect. That said I have no issue posting the 1, 2, and 4 pair versions of your board I have created that have no board modifications otherwise, except for maybe some extra or corrected text.

T6 & T7 will dissipate just over 0.5 watt each with rails of 50v. They do not need to be thermally coupled because they actually perform different functions. But it's easy to use the same heatsink.
Click to expand...

Do not need to be thermally coupled to each other or not thermally coupled to T9/T10?

T9 & T10 perform the same function by driving into the 220 ohm resistor creating the voltage drive for the output FETs. These will dissipate around 0.8 watts each at idle and need to be heatsinked too. They mount on the main bracket because of their proximity to the output devices. Being on the same bracket should keep their temperatures the same and this has some importance with regard to thermal stability once the bias is set. On the other larger module it was too difficult to mount them in this fashion and in practice did not seem to matter as the thermal stability is excellent.


Cheers
Q
Click to expand...

In theory the ideal is for T9/T10 to be thermally coupled to output driver heatsink, which of course is also thermally coupled to T8? If that is case would you feel as one alternative to the creative variations of the PCB I have been creating (with graphics editing) to the TO-247 based PCB to running twisted pig tails to enable mount T9/T10 to be on the main heatsink?


Regards,

John L. Males
Willowdale. Ontario
Canada
11 January 2007 23:05
 
#1,344
Re: Re[15]: Mini - Quasi

keypunch said:
Do not need to be thermally coupled to each other or not thermally coupled to T9/T10?........

In theory the ideal is for T9/T10 to be thermally coupled to output driver heatsink, which of course is also thermally coupled to T8? If that is case would you feel as one alternative to the creative variations of the PCB I have been creating (with graphics editing) to the TO-247 based PCB to running twisted pig tails to enable mount T9/T10 to be on the main heatsink?
Click to expand...


T6 & T7 do not need to be thermally coupled to each other nor to T9 & T10.

Given a choice of extended twisted leads or the current configuration I would choose latter. Connections throughout an amp layout should be kept as short as possible. This is why all my designs (and most others) have drivers and ouput devices on the main PCB. In any case the thermal stability of the original layout is excellent. The only downside is that with the 4 transistors on the one heatsink they get rather warm (hot?). But the MJE340/350's are up to it.

One reason the other layout works well is the seperation of the high current sections from the input and second stage. There is quite a distance in physical terms from the output stage to the input stage, so radiated fields caused by high current are less likely to affect the small signals at the other end. This was not possible with the new mini layout, so it will interesting to see what the measurable affect is.

Cheers
Q
 
#1,348
NMOS200 TO-220 Layout

Quasi,

Nice effort on the TO-220 PCBs. I do not have time to check traces vs schematic, but I have noticed a couple things right off bat that may be helpful.

There is a capacitor just marked "C" near the VPOS fuse.

It might be helpful to mark the GND and SPKR terminals in the PCB to reduce questions that may arise.

Am I blind or is there no inductor for the output on this TO-220 PCB design?

As FYI, my preference of "choice" is the PCB without the extra copper fill.

I would be interested if anyone who has made your TO-247 base design makes this TO-220 design observations on differences sonically and the main heatsink running temperature compared to the TO-247 design.


Regards,

John L. Males
Willowdale, Ontario
Canada
12 January 2007 18:20
12 January 2007 18:24 Typo corrections
 
#1,350
Hi,
4pair of irfp450 running at Tc max=50degC can just manage 380W into 4ohm, IF you can keep the phase of the 4ohm load below 45degrees.
To achieve 50degC at the FET case (Tc) you will need a sink of Rth s-a of 0.11 on each channel.
Tcmax =60degC will require 5pair or take the SOAR to 100mS rather than DC SOA.
 
#1,353
grechejr said:
Anyone ???
Click to expand...

Hi Grechejr,

your answer is: the last table row.
Perhaps the reasoning for this result isn't obvious and needs some explanation...

Quasi came up with some calculated power and output device combinations that he derived from SOA data and voltage/current handling capability of the incorporated devices.

This table of calculated values ends at 85V rails supply and 2Ohm load, because a higher voltage would surpass the rating of some transistors and more current would get the current sharing between the output fets and the overall current consumption out of control.

For your 85V rail supply the table tells us that you could deliver around 350W into an 8Ohm load (using a minimum of 3 fets per rail) or 680W into a 4Ohm load (using a minimum of 5 fets per rail).

Using more output devices (fets) per rail is never forbidden - at least not up to a reasonable and practical limit. More output fets increase the current capability of your amp - and thus the power capacity into a lower load impedance.

But more output fets do not deliver more power into the same load impedance, as this is ultimately limited by supply voltage. For a higher output power into 8Ohm, a larger supply rail voltage would be required ... which the amp just isn't specified for, sorry!

There already have been questions about possibilities to increase voltage supply handling of the amp. Quasi gave some hints in how this could possibly be achieved - given the required knowledge and parts source. He then used to make clear that the amp is specified as is and that he wouldn't investigate into an even more powerful version, as this would essentially change (and overcomplicate) the whole design and - basically - become a whole new amplifier development project.


10 FETS - 85V - 8 Ohm = ___RMS??

what is the delivered power ???
Click to expand...

You name it: around 350W rms (with lots of safety margin to drive lower impedances, too).

Cheers,
Sek.
 
#1,355
Soft Start Circuit Transformer Current

Hi,

I have a question regarding the soft start circuit quasi posted some time back in Post #224:

http://www.diyaudio.com/forums/showthread.php?postid=784687#post784687



I like to know what the 20VAC secondary of the transformer needs to be rated for. I happened across some small transformers that might be suitable for the soft start circuit. I would also be interested to know the 12V curent requirement would be as well.


Regards,

John L. Males
Willowdale, Ontario
Canada
13 January 2007 22:58
 
#1,356
Hi John,

The soft start circuit uses very little power. In practical terms the only load is the relay itself. If the same transformer is also used to power DC detect circuits, plus a few LEDs the total load will still be under 10VA. Just make sure that whatever size you choose, it will deliver 20v AC under full load, otherwise you may find you don't have enough voltage. I suspect that for small transformers the no load voltage coud be as high as 24 or 26 volts.

Cheers
Q
 
#1,357
Hi quasi,

Thanks for your reply.

Is the not more than 10VA current draw you suggested to power one slow start and two DC Protect circuits with one transformer for two modules as would be common for a stereo amp configuration?

The transformers in question are either 20VAC or 16VAC. The former's secondaries are rated at 0.045A (45ma). The store does have some 20VAC 0.060A (60ma) ones, just not enough for the number of amps I want to build. I would use one trasnformer per module using a current rating that would be ok. This means the 20VAC transformers are 0.9VA and 1.2VA respectively, with the 1.2VA ones being in very limited quanity.

These transformers all have seperate secondaries, not CT, that could be paralleled. The voltages would be 10 and 8 respectively, hence the reason for the 12V as I know some builders have used 12VAC (or was it VDC) supply instead. Paralleling the secondaries would double the secondary current of the 20VAC types to 1.8VA and 2.4VA for a 14V secondary. The 16/8VAC secondary transformer is rated at .150A (150ma) per secondary for a 2.4VA rating and 300ma parallel current rating.

All of these transformers do have two seperate primaries for 115/230VAC operation. I assume placing the primaries in parallel would not increase the VA rating at all or is assumed for 115VAC use for the specified VA rating? The 20VAC/45ma I do not know the primary rating on, but the 20VAC/60ma transformer that is very low in quanity has a primary rating of 1.2VA (equals the VA rating of the secondaries in parallel). Oddly the 16/8VAC transformer has a 2VA rating for the primary as opposed to the secondary VA rating calculation of 2.4VA.

I asked about the slow turn only current requirement on as it needs to be a seperate transformer. Some of the main amp transformers I have do have a non CT 12VAC secondary in addition to the main secondary I thought I could use for the DC protect if need be using the lower voltage. For transformers that do not have a supplemental secondary I was going to ask about the DC Protect circuit current draw seperately.

Another option could be to use one of these 20VAC 0.9VA transformers not only for each slow start, but for each DC protect as well. These definitely could be PCB mounted. Using this approach might the 20VAC 0.9VA transformer be of sufficient rating?

I trust you are not close to any of the major bush fires buring "down under".


Regards,

John L. Males
Willowdale, Ontario
Canada
14 January 2007 08:58
14 January 2007 09:13 Typo corrections
 
#1,360
Workhorse.

Are you suggesting the addition of protection diodes in the driver and output stages?

Short circuit protection has been raised and discussed in past. If I recall correctly some of the reasons such protection has not been include it would complicate the design and likely reduce the performance quality of the current design and that the amp was designed for HIFI use in home setting where it is not subjected to the dynamics of field/DJ use of the related system being assembled and disassembled frequently. The fusing is expected to cover the necessary and most reasonable safety concerns.

That said some builders have chosen to use this design in such band, DJ or other similar uses. I would think if they felt a need for such protection they make the required design modifications or add in the additional protection circuit. So far I do not think that has been done by builders using this amplifier in such applications. Quasi does have anotehr design oriented to DJ, field use and higher power in a different thread for those that have a need for such demanding amp applications.

As far as I know none has had any part of this design blow or fry with a successfully and correctly completed effort. Even those that had some mixups in the pinouts of some devices for whatever the reason nothing appeared fo blow with some suggesting some part replacement as being prudent to avoid secondary problems once the correction in the building was made. I think ther was one case of a problem that was likely due to a defective part causing some problems.

However when all was said and done none of the amps built has had any problems. I am nto sure, nor am I sure quasi knows, but my hunch is well over 100 of these amps have been made. I would suggest to you that this is in large part to the design effort of quasi. For sure quasi encourages the use of the DC protection and says so along with many others when a builder suggests they will not use any sort of DC Protection. I also think many of the builders know what they are doing and make sure they are careful in how they wire their systems. Sure we are human and make mistakes. Remember there are no general consumers building and using this design as much as it is likely very robust for the general consumer. So far the quasi design has stood up very well despite some odd and common mistakes in the building phases any amp project. I would suggest to you the design choices quasi made to keep the design simple, have very good part flexibility and the design itself speak for themselves.

Ideally would short circuit protection be desired? Yes. At the cost of reducing the sound quality, robustness, part flexibility and simple design? I do not think so, but I hope others will speak up as well.


Regards,

John L. Males
Willowdale, Ontario
Canada
14 January 2007 09:42
14 January 2007 10:05 Typo correction
14 January 2007 10:07 Typo correction
14 January 2007 10:09 Typo correction
 
You must log in or register to reply here.
Top Bottom