apex b500 merging 4 channels into 2

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi, its been a while since I last posted here, to make it as short as possible the idea is this, ive built 4 channels of the apex b500 years ago, then i built an smps power supply for them and safety circuit etc and put them in a box.
the thing is since I used the toshiba 2sa/2sc1943/5200 devices on them which dont have exactly the highest SOA I would like to simply take the 2 channels that use the toshibas and simply add the drivers to the other two channels.

so that instead of 4 weaker channels I would have 2 stronger ones as they would have more output device pairs in parallel. My psu is quite capable and has output rails of about 85-0-85 so this is one reason I would like to got for this the other one is that two of the 4 channels are built with less quality and they have HUM in them while the other two boards are built with more precision and detail so they are fine.


my question is can I simply merge the outputs and take the signal to them from the active channel driver stage or would the drivers need to be made stronger due to the more output devices?

I have attached the b500 schematic with 3 points labeled 1,2 and 3 of the possible places to connect the extra outputs from the other two channels, what do you think?
 

Attachments

  • schema-diagram-APEX-B500-1024x426.png
    schema-diagram-APEX-B500-1024x426.png
    213.4 KB · Views: 261
well normally one has such ridiculously low loads only in car audio subs etc normally in home or pro audio I see 4 and 8 ohms of impedance for a speaker set.

secondly I did some math , very basic but math and if I take the hFE of the output drivers and then take the worst case scenario current through the load and divide that I get something like this.
my supply voltage is about 85-0-85, which means that each side sees roughly 80v through the load, let's take 80v ad divide by 1.5 ohms we get about 53 amps current. then lets take 53amps and divide by 8 one side output transistors , we get about 6.6amps per transistor.
according to the datasheet of 2sa1943 i get that for Ic 7A the hFE is around 60, so lets take our 53 amps divide by 60 is about 0.88 ,

heres where i get bit confused is this 0.88 or close to 1 amp the total base current for all 8 devices or the individual current per transistor base , I think its the total current for all 8 devices in one side?
 
well basically I have a hard time understanding why many designs especially ones for DIY here which have 7 or 8 pairs of output devices use one of the output transistors in the final driver stage, because according to some simple math I just did here , at near maximum safe levels and very low impedance loads the base current doesn't exceed any significant numbers which a driver transistor like the MJE15032/33 couldn't handle?
Where could there be a situation where a 2sa1943 would need 500+mA of current in its base or say even 1A? because the MJE15032/33 is rated 8A max, so that gives about 1A max per output device if 8 are used. but I think at 1A of base current and a 2 or 1 ohm load the 2sa1943/2sc5200 would have long self destructed into flames from a 80v per side power supply that is strong enough to resist sag.?
 
The driver has less work, if You just add some more output devices, but don't increase the output current.
But if You feel more comfortable with output devices as driver, let's do it! Anyway Fairchild makes FJP5200/FJP1943, which can be Your strong driver...

Sajti
 
Last edited:
yeah but the question is , is there any necessity to use output transistors as the final base drivers if the device which is already there (mje15032/33) can do the job just fine?
the fairchild analogs to the toshiba transistors have a bit higher current ratings but I have toshiba original devices as output transistors.

Well I will do some soldering later this week and then come back with the results basically
 
Member
Joined 2010
Paid Member
....... Where could there be a situation where a 2sa1943 would need 500+mA of current in its base or say even 1A? because the MJE15032/33 is rated 8A max, so that gives about 1A max per output device if 8 are used. but I think at 1A of base current and a 2 or 1 ohm load the 2sa1943/2sc5200 would have long self destructed into flames from a 80v per side power supply that is strong enough to resist sag.?
As you say, you have looked at the loads and calculated simply - assuming they are only resistive. However, speakers are reactive loads and at the sort of power you seem to need, the peak base currents of even lightweight (150W) devices like 2SC5200/A19434 will be significantly higher than the datashheet Hfe figure might suggest.

Unfortunately, estimating load lines and peak power requirements for output transistors is not a simple, steady state DC calculation covered by Ohm's Law. You need to search load lines, DC and AC, for class AB audio power amplifiers to get a better understanding. As it is an engineering level topic, it may require a bit of heavy mathematics. I found this paper gave a fairly clear overview: http://highered.mheducation.com/sites/dl/free/0070634246/329094/malvinoch12.pdf
 
Ok I will read , I understand speakers are reactive , and so I was thinking that maybe using mje15032/33 could introduce some distortion upon heavy loading if at some point the driver exceeds it capability to power the bases of the output trannies.

although when I calculated using the hFE numbers mentioned for given Ic in the datasheets I used a load assumption which would go far lower thna I would ever play it, because normal home audio or pro speaker have 4 to 8 Ohms , I assume 1.5 ohms of load at 80v of voltage per side since my power supply is 85-0-85
at these numbers I got a worst case output current of 53 amps through each sides 8 output devices.
So I split the 53 amps upon my 8 devices and get about 6.6 amps per device which is very close to the toshiba datasheet where 7A is mentioned and a typical hFE is given to be around 60 at that collector current.
So then I simply divide the 7 amps with the hFE and get about 111mA of base current but even assuming a much higher base current is still within reach of the MJE15032/33 Ic of 8A.

Isn't this approach reasonable given that I used impedances much lower than would be actually used?

The one thing I'm thinking about is whether I would need to up the bias current in the driver stage because additional output could mean that the bias could drop below a level where the transistors start to close upon certain conditions and introduce "zero crossing" distortion.
 
Member
Joined 2010
Paid Member
Don't assume that there will be equal current sharing between the output devices, as you would need equal die temperatures and close matched sets of transistors to stay within about an 8% current sharing error range and that's unlikely - you may need more safety margin unless you have a good matching jig and test set-up. Why not ask Apexaudio for advice on how he would go about driving 8 Prs of output transistors in the B500 design?
 

PRR

Member
Joined 2003
Paid Member
It is "simpler" to use an extra "output" device as a driver. Your workers don't put the wrong part in the wrong place. You need half as many inventory bins. You may get a better price on a larger number of outputs than a small number of small drivers.

Yes, an output device as the driver is "over-qualified". Like the small shack I am buying, which has a massive foundation. That's not necessarily bad.

85V rail each side. Simple resistor math says about 10 Amps peak. With four outputs, about 2.5A each; with eight, 1.25A each. (Though sharing is uncertain.)

> I think its the total current for all 8 devices

Yes. I think "1.5 ohms" is awful low. I figure 10A peak, min hFE 20, half Amp total base current. Assuming hFE does not vary much between 1.25A and 2.5A, eight or four devices take about the same base current. Perhaps the hFE is a hair higher with 1.25A, so the eight-device scheme will use a hair less base current.

Yes, scaling to 1.5 Ohms and hFE>50 does give around 1 Amp total. At this high current I would look closer at the hFE curve. Many devices will be falling hFE at 7 Amps, so the 8-part final may be noticeably less base current. Back when hFE was failling by 1 Amp, we sometimes paralleled for better current gain rather than dissipation or collector current limit.

However 7 Amps peak would worry me about SOA. Until recently few parts could stand much over 1A at 40V. On resistor you could swing 80V or 2A peak, but oval load-lines would cross the SOA line and melt your bits.

IMHO, commercial hard-duty (PA) amps got a lot more reliable when the price of transistors fell so low a designer could reasonably use a dozen devices instead of four. Kill an original Phase Linear 400 (not hard) then go pop the hood of a modern PA-market equivalent: long rows of final devices.
 
Last edited:
well the problem is APEX never responds via PM's I've tried years ago and a few days back - the same, as for the amplifier , I did join the extra transistors and merged the channels and now I have two channels running 16 output devices each on 80-0-80v rails and as far as listening experience I would say nothing has changed, sadly I dont have large enough speakers to test this amp to its full ability in a reactive load with some scope probes attached but i'm sure I will do that or the same into a resistive dummy load some time in the future.

ill post some info or pics if anyones interested later as I have them on my phone but write from pc.


what do you think could this extra output transistor while using the same driver could influence the sonics and frequency response of the amp?

I noticed the bias current is now lower since it is shared among twice as much transistors as before but the goodside is that the heatsinks say colder as before they got i'd say too warm at idle with time.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.