AndrewT said:
Ahh yes, you are right. I was thinking undriven, there is only about a 1.9V level at the base of the VAS, which has a total of 390 ohms emitter resistance. So the VAS could only sink 30ma, even shorted to ground by the protection circuit. Well, perhaps short term that could be borne, with good heatsinking.
But driven, there is about a 3.8V peak drive there, which would allow the VAS to sink much more, say 80ma. It would fry. That's the point I missed. Thanks!
I'll have to look at the current limit, but I think you are right, his 33R may allow too much current to flow, depending on the heatsinking provided for the VAS. I suppose it is a tradeoff between conservative protection and allowing plenty of VAS current. But really an output triple needs low drive current, doesn't it?? Hmmm... Have to agree with you there too.
But I guess once again the topology of the amp is vindicated. We can tweak values, but the right stuff is there. Another tip of the hat to Dr. Leach. And legions of students who beat the dickens out of the design to get the bugs out....
Hi,
theoretically the VAS is operating in ClassA and swings either side of Iq, so the maximum current is about 2 times Iq.
To allow for asymmetry (if it exists) and to exclude the risk of early protection tripping, I believe that a trip target of 3 times Iq should mean that the VAS will never go into protection (CCS) mode into ANY normal speaker load.
That's where my 56r came from, allowing VAS Iq=3.6ma
Many other amps have a VAS protection in there, but I wonder if any predate Leach from 1976?
theoretically the VAS is operating in ClassA and swings either side of Iq, so the maximum current is about 2 times Iq.
To allow for asymmetry (if it exists) and to exclude the risk of early protection tripping, I believe that a trip target of 3 times Iq should mean that the VAS will never go into protection (CCS) mode into ANY normal speaker load.
That's where my 56r came from, allowing VAS Iq=3.6ma
Many other amps have a VAS protection in there, but I wonder if any predate Leach from 1976?
Hi,
the VAS transistors chosen by Leach are 1W.
On the bassis of choosing a 40Vac transformer and possibly holding 58Vdc on the rail during protection triggering the maximum current the VAS can sustain medium term is 1000/58=17.2mA.
Leach's chosen sensor resistor allows about 20mA (varies depending on trigger voltage) to pass, effectively restricting the safe Vce to just 50V, well short of full rail voltage. This may work if the rail is severely drooped and peak fault current flows for only a short time. I think 20mA is sailing too close to the wind.
Has anyone a thought on the necessary ratio of VAS Iq to pre-trigger fault current? Is 3:1 about right. Could we mange with much less say 2:1 or even 1.5:1 or should we be aiming for closer to Leach's figure of 5:1?
Is there a way to test the amp to check at what level the Vas protection activates?
The last thing I want is someone coming back saying I built it your way & I can hear the VAS protection trigger activating.
edit:- OOPs, my numbers are referring to the Leach low tim 4.5
the VAS transistors chosen by Leach are 1W.
On the bassis of choosing a 40Vac transformer and possibly holding 58Vdc on the rail during protection triggering the maximum current the VAS can sustain medium term is 1000/58=17.2mA.
Leach's chosen sensor resistor allows about 20mA (varies depending on trigger voltage) to pass, effectively restricting the safe Vce to just 50V, well short of full rail voltage. This may work if the rail is severely drooped and peak fault current flows for only a short time. I think 20mA is sailing too close to the wind.
Has anyone a thought on the necessary ratio of VAS Iq to pre-trigger fault current? Is 3:1 about right. Could we mange with much less say 2:1 or even 1.5:1 or should we be aiming for closer to Leach's figure of 5:1?
Is there a way to test the amp to check at what level the Vas protection activates?
The last thing I want is someone coming back saying I built it your way & I can hear the VAS protection trigger activating.
edit:- OOPs, my numbers are referring to the Leach low tim 4.5
In my experience the current limit activates only when driving about 6dB into clipping with a 2 ohm reactive load (a pair of 4 ohm 15" woofers in parallel, 3 pair of MJ15011/12 on ±63V rails.
The 2N3440/5415 are rated at 10W at 25°C vs the MJE340/350 being 20W. With the junction running at 100°C they are both rated at the same power as the metal is a 200°C part vs 150°C for the plastic.
If you are worried, you can always run a heatsink.
The 2N3440/5415 are rated at 10W at 25°C vs the MJE340/350 being 20W. With the junction running at 100°C they are both rated at the same power as the metal is a 200°C part vs 150°C for the plastic.
If you are worried, you can always run a heatsink.
Hi Djk,
are you referring to Leach or super Leach in the current limiting example?
are you referring to Leach or super Leach in the current limiting example?
does not compute for super Leach, it must be even number of pairs.
It was a regular Leach, circa 1980, with three parallel pairs of the 10A rated MJ15011/12.
The point was that the Vas doesn't go into current limit on any kind of normal load or program material, only when the output foldback current limiter kicks in. The amp is grossly overdriven at that point, and the output limiter causes the amp/speaker to make loud popping noises from the inductive load in flyback.
The point was that the Vas doesn't go into current limit on any kind of normal load or program material, only when the output foldback current limiter kicks in. The amp is grossly overdriven at that point, and the output limiter causes the amp/speaker to make loud popping noises from the inductive load in flyback.
Hi Djk,
yes, I agree the Leach VAS current limit is controlled by the 30r in the VAS emitter connection. This value allows about 5times the Iq to flow before limiting sets in. I consider this to be excessively high. However for the 1000mW transistors normally specified here that is just about bearable for a shortish term overload. Pd is about 50V * 20mA until the junction starts to heat up and then the transistor is going to fail if the overload condition continues. These small transistors will not survive very long due to the small heat inertia that is inherent in their manufacture.
I would like to suggest reducing the limiting current to a much safer value. Theoretically it only needs to be a maximum of 2times Iq and possibly significantly less. But I fear that reducing it this far by adjusting the 30r up to 75r (8mA) may make the limiting audible when a lower value (47r to 56r) might suffice. That's where my 3times came from.
But is 12mA a good compromise that is inaudible and provides effective protection for the VAS?
yes, I agree the Leach VAS current limit is controlled by the 30r in the VAS emitter connection. This value allows about 5times the Iq to flow before limiting sets in. I consider this to be excessively high. However for the 1000mW transistors normally specified here that is just about bearable for a shortish term overload. Pd is about 50V * 20mA until the junction starts to heat up and then the transistor is going to fail if the overload condition continues. These small transistors will not survive very long due to the small heat inertia that is inherent in their manufacture.
I would like to suggest reducing the limiting current to a much safer value. Theoretically it only needs to be a maximum of 2times Iq and possibly significantly less. But I fear that reducing it this far by adjusting the 30r up to 75r (8mA) may make the limiting audible when a lower value (47r to 56r) might suffice. That's where my 3times came from.
But is 12mA a good compromise that is inaudible and provides effective protection for the VAS?
For a fault from overdrive with music, it's OK. If the output stage blows, the secondary failure of the Vas is of little importance. It won't hurt to over-design it.
Does anybody have a complete construction manual of it? Anyone at all? Or had it never been finished?
Tks
Tks
Jens is falling out as I can understand from that post. Is anybody else willing to show their experience on it?
bigpanda said:
I have not fallen out of the project; I just run it for myself......
Current layout status is that all parts are in place I need to clean the desing up.
This is not going to be public domain so please don't ask for boards or gerbers.
\Jens
Jens, hope that I did not offend you. I have got everything ready after months of prep. I am just seeking help from anyone who might be willing to do so for the pre-freight test...... and that kind of stuff. I have been thru with the boards (only if there has been changes again)
alternative for output transistors
Hi,
Sorry to dig this up again. I order some MJL4302A but never get them. I happen to have some MJL21193, MJL21194 at hand. (I don't expect them to work as powerful as the original) Can they be use instead? Any modification of circuit need for using them? Or had it been mentioned in the thread and I was just blind enough to have missed it?
Tks for your help.
Hi,
Sorry to dig this up again. I order some MJL4302A but never get them. I happen to have some MJL21193, MJL21194 at hand. (I don't expect them to work as powerful as the original) Can they be use instead? Any modification of circuit need for using them? Or had it been mentioned in the thread and I was just blind enough to have missed it?
Tks for your help.
Hi,
are 4302A the lead (Pb) type?
You might only be able to get the lead free (no Pb) 4302G?
Were you talking samples from ONsemi? They tend to restrict samples to the newer parts.
Maybe they expect us to research the new product lines before we order up thousands of components for the latest specification production run.
MJL21193/4 is a pretty robust output pair. It should do the job very well. It is slightly down on power but this is of little consequence.
21193/4 does a lot better than 4302 when looking at 50V and 70V current ratings. Cascoded 21193/4 running on +-100Vrails will only see 50V each when driving maximum current into a highly reactive load. Here the 21193/4 will beat the pants off a 4302 with nearly double the current capacity.
ps send all your As to me. I don't care what goes inside my equipment, it never gets dumped!
are 4302A the lead (Pb) type?
You might only be able to get the lead free (no Pb) 4302G?
Were you talking samples from ONsemi? They tend to restrict samples to the newer parts.
Maybe they expect us to research the new product lines before we order up thousands of components for the latest specification production run.
MJL21193/4 is a pretty robust output pair. It should do the job very well. It is slightly down on power but this is of little consequence.
21193/4 does a lot better than 4302 when looking at 50V and 70V current ratings. Cascoded 21193/4 running on +-100Vrails will only see 50V each when driving maximum current into a highly reactive load. Here the 21193/4 will beat the pants off a 4302 with nearly double the current capacity.
ps send all your As to me. I don't care what goes inside my equipment, it never gets dumped!
- Status
- Not open for further replies.