Smaller Leach Amp V1

Status
Not open for further replies.
Jump to Latest
#821
jacco vermeulen said:
Terry,
taking jens's circuit :

for your 64 volts(70 volts max) :

R28/R30 = 270 Ohms, as original leach

For 0.33 Ohms emitter resistors:

R41 through R62 = 620 ohms, or next best value.
R26/R34 = 3K9 as original Leach

For 0.47 Ohms emitter resistors:

R41 through R62 = 1K2
R26/R34 = 5K1

With these values a 6-device Leach on 64 vdc rails can do about 10 amps continuous and over 40 amps peak.
Under normal circumstances without peculiar drivers those currents are so high the protection will never be triggered, sound quality will not suffer.
And you have a solid safety margin, that professor knows his business.
Click to expand...


Ok, That's clear enough.

Any advantage either way of using 0.33 vs 0.47 for the emitters or should I just pick one?

Thanks again, Terry
 
#823
I think Stuart Easson posted plenty enough on emitter resistors on the Krell thread.
Sort of depends on the quality of the resistors, imo.
With Stuart's meanderband 0.68 Ohms i am pleased to employ such high resistance, for both the Krell as the Leach.
The higher the resistance the higher the wattage you need to employ.
A 0.33 ohm resistor handles more current than a 0.47 Ohms resistor of the same wattage.
 
#824
Hi Jacco,
are your calculations on protection values based on the 230w/250w heat dissipation of the output transistors?
If they are then you need to take account of the rise in case temperature since this reduces the allowable SOAR. The case temp will depend on the bias current and heatsink and on the recent power output history, this being dependant on speaker load impedance and delivered output power. ClassAB transistor temps rise as output power rises.
Under reasonable domestic loading the SOAR could reduce by between 10% and 30% due to case temperature considerations only.
The next part of the SOAR reduction is due to the second breakdown region of the BJTs which for many of the recommended BJTs for the Leach start to reduce with Vce above 40V to 50V and are down to half power by about 70V to 80V. This is made worse by a reactive load requiring current when the output voltage is near (above or below) the zero level.
Combining second breakdown with case temp could reduce permissible current to less than half the 25degC power levels asummed in my first question.
The writer that suggested that the protection is bullit proof could be well short if he reviews the actual working SOAR after accounting for temp & voltage. It will become bullit proof if the resistor values are adjusted to match the actual amplifier (or worst case) conditions.
The caps C17 & C18 have a very big influence on delaying protection activation even if you opt for the safe value resistors permitting only half power dissipation in continuous duty. The output currents could approach and even exceed the 25degC rating when they are of short term peaks only (pretty typical of music).
Finally, thanks for that expose on calculating the actual resistor values and explanation of how to get there for different amp setups.
 
#825
Here are my calculations for Jen's circuit using the MJL3281/MJL1302 pair with 56V rails and 0.47R emitter resistors:

R28/R30

270R

Current through R28 = 0.6V / 270R ~ 2.22 mA

R41/R42/R43 and R60/R61/R62

200W / 61.5V ~ 3.25A

3.25A * 0.47R ~ 1.53V

1.53V - 0.6V = 0.93V

0.93V / 2.22mA ~ 420R

3 * 420R = 1.26k --> 1.2k

R26/R34

15A Max Rating --> 13.5A

13.5A * 0.47R ~ 6.35V

6.35V - 0.6V = 5.75V

5.75V / 1.2k ~ 4.8mA per emitter --> 14.4mA

14.4mA - 2.22mA ~ 12.2mA

61.5V - 6.35V = 55.15V

55.15V - 1.3V (for Vbe and diode drop) ~ 54V

54V / 12.2mA = 4.5k --> 4.3k

I figured I'd post these because there might be others looking to use the same devices and to also make sure my math is correct. Thanks!
 
#826
Actually i just took a look at the current limit graph of Prof Leach's homepage and checked the components of the VI limiter with the basic Ohm rule if it corresponded with the zones in the graph.

A number of people on this forum posted they agree that a good way to cope with SOA areas is to derate these by 50% for realistic temperature conditions and safety reasons.
Derated to 50%of the SOA curve, at 87.5C for devices with a die max temperature of 150C, seems to be a practical choice that has sufficient safety reserves and at the same time retains enough of the power capability of the device.
Nearly all designs i have seen seem to be governed by that rule.
I have been calculating current capabillity of amplifiers as a sidekick since i started this hobby, based on; rail voltages, number and type of output devices, insulation material, device case type, emitter resistor value.
Seems even the good Dr Leach follows the rule.
Up till a few years ago most of the circuits i saw had a rigid current protection, as good as all of the designs in electronics magazines.
The ones that had exotic protection circuitry usely had a devistating effect on the quality of the amplifier.
Surely a great number of parameters influence the current abillity of an amplifier at any moment; characteristics of the load, ambient temperature, source material, besides the ones that are steady state such as type of device case, heatsinking etc.
Funny thing is that the 50% soa rule seems to be/have been applied no matter the class of amplification, type of output device, or casing type.

I think it is safe to assume that the VI limiter of the V4.5 Leach functions properly, the amplifier has been around long enough and i can only guess how many built it.
If basic electronics applies for the VI limiter of the V4.5, why should it not for Jens's V5.8(.10 ?)

With a more intelligent current limiter much more could made from the soa, Mikeks showed so with his triple locus limiter.
If one, Michael might be the one to see which values suit the VI limiter components best.
I merely posted how the VI limiter of Dr Leach works and how values could be altered following the same line of thinking.
 
#831
Look for a multi-turn pot for better control/resolution. Also, a conductive plastic one will have lower noise and better reliablilty than, say, a carbon one, if they are in fact still made.

Another trick is to get a panel mount one with leads that you can attach to the board. This way, you can adjust bias outside of the enclosure. Really beneficial if you put pin jacks on the outside of the case attached to the emitter resistors, so you can simply stick your DMM probes into the pin jacks while you adjust the bias based on emitter voltages, all from outside the enclosure.

For the really deranged DIYers, Vishay makes a bulk foil trim pot that is case mountable.

See, for example, the 1202LB model at the bottom of this link:

Vishay pots
 
#832
I just got out the transformer that I am planning on using. It measures 48.4VAC to the CT. If my calcs are correct that means 68.44VDC less rectification. I'm probably going to be a little higher than the 64V rails that I had quoted. Is this going to throw off the math that jacco so graciously did for me?

I also checked the other winding that I had planned to run the front end with. It measures 55-0-55 VAC. I don't think it has the same vA as the other winding. Anyway it would end up with 77V rails. Probably way too high to run the front end with only 48V rails on the outputs. Is this right?


Thanks, Terry
 
#833
Terry,

Remember that the output stage voltage will probably drop a couple more volts when loaded. My 1 KVA transformer put out 65 volts rectified and filtered, but drops to 61 under the load of four idling Leach channels. So:

A. jacco's calc are close enough, even at 68 V. (he added 10% or so for line fluctuation)

B. 77 volts will also probably drop a couple volts, depending on the winding size. Probably still more "extra" than you need, but it does give you lots of room for a regulator to work. Regulators have been thrashed to death in other threads, but you have many options.

Since you can make boards, why not use the regulator used in the Pass/Thagard A-75 using the layout provided in the article? Substituting a 12V zener for the 9.1V reference gives about 64.5 volts. If you want more you can fiddle with the voltage divider resistors or add a forward biased 1n4007 in series with the 12v zener(giving a reference voltage of 12.6V) to get you around 68V. (the output is 5.5x reference voltage in stock form) I've built several versions of this regulator on perfboard with good results if you don't want to be bothered making boards.

Edit: if you use this regulator, don't forget to upgrade the output caps to at least 80V. You won't need the voltage doubler section.
 
#836
still4given said:
Is this right?
Click to expand...

Terry,

you have build enough amplifiers till now to see that Prof. Leach made the V4.5 a very reliable design.
For the protection circuit he must have taken the worst case scenario, with 10% higher voltage coming from the powerplant and supposing that the output devices can reach a temperature that reduces the safe operating area to halve of that at 25C.
Make sure that your Leach amplifier stays cool enough by being generous with the heatsink, and with your rail voltages you still have plenty reserve to keep it alive.
The Krell you built used the same type and number of output devices.
It has much lower rail voltages, but runs so much hotter, and it did not even have a current limiter.
When the amplifier is still cool, such as when it is just switched on, the current capability is so much higher than the protection limit.
In the case that you intend to use it as a subwoofer amplifier i would pay more attention to giving it an oversized heatsink instead.

Peter Daniel posted he built a number of A75 regulators.
Some time ago he put a set he made up for sale, posted pictures of them. He said that the regulator did a good job, looks like a good suggestion for your setup.
 
#837
jacco vermeulen said:


Terry,

you have build enough amplifiers till now to see that Prof. Leach made the V4.5 a very reliable design.
For the protection circuit he must have taken the worst case scenario, with 10% higher voltage coming from the powerplant and supposing that the output devices can reach a temperature that reduces the safe operating area to halve of that at 25C.
Make sure that your Leach amplifier stays cool enough by being generous with the heatsink, and with your rail voltages you still have plenty reserve to keep it alive.
The Krell you built used the same type and number of output devices.
It has much lower rail voltages, but runs so much hotter, and it did not even have a current limiter.
When the amplifier is still cool, such as when it is just switched on, the current capability is so much higher than the protection limit.
In the case that you intend to use it as a subwoofer amplifier i would pay more attention to giving it an oversized heatsink instead.

Peter Daniel posted he built a number of A75 regulators.
Some time ago he put a set he made up for sale, posted pictures of them. He said that the regulator did a good job, looks like a good suggestion for your setup.
Click to expand...


Hi jacco,

Since this is mostly for a learning experience I would like to learn about regulated supplies. Do you guys have a link for the foil pattern and/or schematic for this circuit?

One question, Does the front end need as much current as the outputs? I don't think the second circuit has as much vA as the lower voltage winding. I'm not positive of that but the wires are smaller on the higher voltage winding.

Thanks again, Terry
 
#839
You can see how the transformer I want to use is set up in the Hafler XL-280 here. Look at page 17 of the manual. On the voltage chart on page 18 it shows the one winding at +/-75 and the other at +/-65. I suppose that's what I should expect.

Looks like it uses 2 X 470 uf 100v filters for the front end and 4 X 7800 uf 75V for the outputs. I guess that shows the the front end doesn't need as much current.

Thanks, Terry
 
Status
Not open for further replies.
Top Bottom