Project 11.1 from Slone "High-Power Amplifier" Book

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Does anyone have experience with the "11.1" project from the Randy Slone book, "High-Power Audio Amplifier Construction Manual"? It is the first and simplest amplifier in Ch. 11. What is the wattage? Can it handle a minimum load of 2.5 Ohms presented by an e-stat? Slone says early on that all of the amps in the book will drive a 2 Ohm load, but only temprarily. The nominal load is 6 Ohms.

BTW, I love the book. I like his approach and design philosophy. I am anxious to get building.

Regards,

Karl Lewis
 
Karl71 said:
Does anyone have experience with the "11.1" project from the Randy Slone book, "High-Power Audio Amplifier Construction Manual"? It is the first and simplest amplifier in Ch. 11. What is the wattage? Can it handle a minimum load of 2.5 Ohms presented by an e-stat? Slone says early on that all of the amps in the book will drive a 2 Ohm load, but only temprarily. The nominal load is 6 Ohms.

BTW, I love the book. I like his approach and design philosophy. I am anxious to get building.

Regards,

Karl Lewis

Most amps will drive a 2 ohm load so long as you dont turn them up too far. Too low a laod takes too amny amps from the amp power supply before max volts is reached.

I often play my amp into 1ohm but dont turn it right up.
 
Hi!

Not sure if this is the amplifier book I remember with that sort of name or not... if it is the same one, avoid! Otoh, it may be a different one. :)

Perhaps if you post the schematic, people could make intelligent comments, or unintelligent comments about it.

Generally speaking, if the driver can handle enough current, you can increase the number of outpoot devices, and so be able to run more current into your load.

In order to run a lower Z load you need to run more current into the load. Take the power supply rail and figure out what the current requirement into the load impedance will be, then work backwards to look at that SOA of the outputs... then you will know the answer to your question for any specific amplifier.


_-_-bear
 
The 11.1 is a simple amp intended as a beginners project. It has 50 V rails and a single pair of 2SJ162/2SK1058. The max Id rating of those is 7 A, which you reach already at 14 V output for a 2 Ohm load. Obviosly you could reach several times the max Id with that rail voltage. You could lower the rail voltage to maybe 20 V (which is probably safe considering Vgs and other losses) or use parallel output devices as Bear suggested, but it is hardly worth it for a simple amp that not even the author thinks is very good.
 
Nigel -

That's what I have read. I don't intend to run the 'stats very loud. The 2.5 Ohm impendence occurs at around 12 KHz.

Bear -

Why should I avoid the book? I've read past posts regarding Slone's works, and I know that there are conflicting opinions regarding his positions and designs. What are your objections?

Christer -

So, do you think I could use the 11.1 amp with the 'stats? The builder told me that Rotels, Adcoms, and NAD units have had no trouble with his panels. Would a higher power design in the Slone book be more appropriate in this application?

Thanks to everyone so far for your imput. Keep it coming.

Karl
 
"max Id rating of those is 7 A, which you reach already at 14 V output for a 2 Ohm load. "

Which is 10V RMS for a whopping 50W at 2 ohms.

For better results, reduce the main supply voltage to like ±30V and use a higher voltage like ±40V for the front end. It still won't put out any more power, but now it won't be a space heater.
 
"And even 30 V rails are enough to make the amp suicidal with a 2 Ohm load,"

If he was driving a regular speaker I might agree with you, and drop the main rail even lower.

The E-stats only drop that low at very high frequecies. With the low program content and high peak-to-avereage ratio up there I think he may be safe.

" especially since 11.1 is intendend as a very simple project and doesn't have current limiting or any other form of protection."

An output fuse and a thermal switch on the heatsink are generally sufficient for lateral MOSFETS.

The Hafler XL-280 used three pair of outputs on ±63V main rails with a higher voltage for the front end, it drove 2 ohms quite well. If we cut the ±V in half the dissipation will be cut by about 75%, so one pair ought to do (with his speakers).
 
I said, IF that is the book I recall...

post the schematic, is what I said??

Anyhow, it is usually difficult to make a silk purse out of a sow's ear?

A better idea might be to start with a project that is known to work for your application? That is one that has sufficient current capability for the power level that you want to run at?

In general terms, ESLs are fabulously low sensitivity speakers, which means in practice that you want to throw the maximum voltage swing you can muster at them.

IF you have an ESL with a dip down to 2 ohms, then you might want to reconsider how you match to them? That might be the reason that you have such a dip? The place for that is in the ESL section here... :D

But, assuming you are going to need to drive the 2 ohm part, as has been mentioned it only is a big problem if you have to put significant energy there for a while - in terms of blowing up the amp - but in terms of practical listening, you will not like the sound of your amp clipping at all when you run an ESL!! Very annoying, imho.

That's one reason people prefer tubes on ESLs in some situations!

Anyhow there are a number of amps that have been suggested on DIYaudio where people have made up PCBs for the driver part, and there are the Pass designs - some of which will drive 2 ohms all day - and there are known commercial kits that range from blank PCBs to full parts + PCBs and instructions.

For a first ever building project, I'd suggest staying within the range of something that is well known and relatively easy to put the parts together so that it actually works.

The JC-3 amp thread might be worth a look too - or was the JC-2? I get the nomenclature mixed up. Someone made a PCB up for that, and it is a Class A amp that should not mind the low Z load, assuming the proper number of outpoot devices...

These are just some ideas...

Oh, there are also some Opamp based amplifier projects that might work just fine for you! There's an app note that uses a new ultra low distortion opamp as a front end for a power amp: National Semiconductor App Note AN-1645 you probably also want to look at AN-1490 which is referenced in the first one.

No reason that I see that you can't bridge a pair and get substantial voltage swing... and maybe even bias them up into near Class A...

So, P = I^2 R

100 = I^2 2ohms

100/2 = 50 = I^2

= 7 amps


BUT, you want 100 watts (for example) at 8 ohms!
That works out to be 200 @ 4 and then 400w @ 2 ohms!

So let's refigure:

400/2 = I^2 = 14 amps.

Since about a 200w @ 8 ohm amp seems to have about the right voltage swing to run most ESLs, that means even more current available! (whoa!) (but the good part is that some transistors today have ample current and voltage ratings)

Now in practical terms, you're not going to actually pull this current running an ESL, but the ouput devices need to have the SOA so that they do not blow up in use - the ESL looks like a capacitor as a load. (rather reactive, not resistive)

So, the amp also needs to be stable into this load. (important!)

The reason that your ESL is dipping down to 2 ohms is that there is some combination of factors, first the reduction of Z as frequency goes up, and probably in this case (if it dips and comes back up) a resonance of the secondary of the ESL driver transformer with the panel itself (or some other component added in...) causing an impedance dip, and doubtless a phase shift WRT to frequency.

You want an amp that will not blow up or clip before you reach normal listening levels WRT the power level in the majority of the spectrum - which works out to be the voltage swing required when run through the ESL's transformer necessary to deflect the diaphragm sufficiently (on the order of 5kv!). Ok then?

Put another way, to run most ESLs you want an amp with an ample SOA, and that means current capability, PLUS the most voltage swing you can manage - that equation usually means a big, high power amplifier in practice.

_-_-bear
 
Member
Joined 2002
Paid Member
Karl71 said:
Does anyone have experience with the "11.1" project from the Randy Slone book, "High-Power Audio Amplifier Construction Manual"? It is the first and simplest amplifier in Ch. 11. What is the wattage? Can it handle a minimum load of 2.5 Ohms presented by an e-stat? Slone says early on that all of the amps in the book will drive a 2 Ohm load, but only temprarily. The nominal load is 6 Ohms.

BTW, I love the book. I like his approach and design philosophy. I am anxious to get building.

Regards,

Karl Lewis

Hi Karl,

This amp is very similar to the Maplin 100W MOSFET amplifier (55 volt rails) and I think you'll find it's a classic design that pops up quite regularly.

http://www.cpemma.co.uk/map_100w.html

I like this book as well but some of the criticisms are quite valid. Keep a close eye on the schematics as there are a few minor and some not some minor errors (or inconsistencies) as mentioned in a previous post by Nuuk.

For example, if you compare the schematic of 11.1 and the PCB (figure C.1) you'll find the location of the fuses is different. In schematics 11.8 and 11.9 the value of R2 seems wrong.

regards
 

taj

diyAudio Member
Joined 2005
bear said:

post the schematic, is what I said??



Bear,

It's a published book, currently available for $. We must respect the intellectual property (and the copyright that "protects" it), and not post his schematics.

I know that's tough to comprehend for those not personally involved in producing intellectual property and depending on its income.

..Todd
 
Member
Joined 2002
Paid Member
taj said:



Bear,

It's a published book, currently available for $. We must respect the intellectual property (and the copyright that "protects" it), and not post his schematics.

I know that's tough to comprehend for those not personally involved in producing intellectual property and depending on its income.

..Todd

Hi Todd,

I think it has been discussed before, but it is my understanding that the copyright is only the printed material, not the actual schematic. So you can't photocopy it or scan it but redrawing the schematic is OK.

In this case, it can't really be considered an exclusive Slone design as it is almost the Maplin design and that probably came from someone else anyway.

regards
 
Greg Erskine said:


Hi Todd,

I think it has been discussed before, but it is my understanding that the copyright is only the printed material, not the actual schematic. So you can't photocopy it or scan it but redrawing the schematic is OK.

In this case, it can't really be considered an exclusive Slone design as it is almost the Maplin design and that probably came from someone else anyway.

regards


I bought the Maplin 100W amp and was disappointed with it.
The power output was poor from it.
In the end I went for the Maplin 225WRMS amp and that was good. I coupled it to four Fane 50WRMS speakers an it was very loud.
 
As I understand it, under the "fair use" doctrine/clause one can excerpt sections of copyrighted material and put them up on the internet.

Certainly one could do so for say "half" of the schematic.
Then post the other half...

Or perhaps, if ur worried, remove the parts values, and just publish the topology...

Redrawing is certainly an option...

Dunno what the law is elsewhere, but that is my understanding for the USA and probably all of North America...

_-_-bear
 
bear said:


Dunno what the law is elsewhere, but that is my understanding for the USA and probably all of North America...

And there came the most interesting issue, which laws to follow? Does it depend on where the server is located? Or is it the poster who takes risks if violating local laws in his/her country? Can readers violate the law in any country by just looking at the schematic, or at least by saving it as a file?
 
Ok,

I looked at the schematics...

A few comments.

Clearly one is not going to be able to increase the number of output devices with this schematic. So, that leaves it out for driving a 2 ohm load of any sort, with the possible exception of a very low power speaker - not an ESL.

And, it is a fairly horrid example of a very basic amplifier design - one that in essence counts on the global feedback to make acceptable the performance, since without the global feedback the thing would be all over the map in terms of results - meaning the lack of matching of devices would be hyper evident.

It would be nice to enforce or create true balance and differential operation of the long tail pairs as a starter point. The lack of a decent Vbe multiplier at minimum for bias setting is pretty outrageous this being 2008...

OTOH! If you've never built an amp, and this is your first pass, any amp that you build and make work is wonderful!! :D :D

But, it still won't work right on the ESL.

That IC from National might work out for you... it eliminates all the guesswork, and the matching issues for the front end (if any), and leaves you building a driver stage and output stage + a perf board with a few parts for the driver IC... not too shabby for a first project!

See what they say in the 1490 app note about driving 4-6 Mosfets per rail... that's ur goal... that and a big **** heatsink to go with it. you can build the whole amp onto the heatsink I think, except for the power supply.

_-_-bear
 
Bear, don't bash Slone to hard. This amp is the first of twelve in his book and is intended as a simple beginners project. He points out in the book that it has problems, lika rising distorsion at high frequenicies. It isn't intended to be good, but simple. I think all the other 11 amps have both some form of current or SOA limiter and Vbe multiplier (even the others with lateral MOSFETs, which may not strictly need it). Elsewhere in the book he also shows an elaborate speaker protection circuit, which is the only one I have seen that even tries to detect amp oscillation!
 
"It would be nice to enforce or create true balance and differential operation of the long tail pairs as a starter point. The lack of a decent Vbe multiplier at minimum for bias setting is pretty outrageous this being 2008..."

Laterals don't need a vbe multiplier. At 100mA Id the temperature coefficient of the current is zero. The bias will vary with + rail voltage though as LTP tail current sets the bias current through the VAS and bias pot too. A vbe multiplier or current source for the diff pair would make the bias more stable with rail voltage.

Not having gate resistors on the FETs looks a bit scary though, it might oscillate if the lauout is not perfect.

For protection: Add 10V zeners in series with diodes to limit gate voltage and you will have a very hard time destroying those FETs.

I tried to destroy a lateral fet by mounting it on a very undersized heatsink, connecting 50V drain to source and apply 10V current limited to the gate. The transistor did not blow up - when it got overheated some kind of internal thyristor structure shorts gate-source turning the transistor off until gate current is removed.

One pair driving 2 ohms won't give much power though before current limiting, but if this just is the impedance at the highest frequencies you will probably be fine unless you like high power 20kHz sinewaves :p

"Clearly one is not going to be able to increase the number of output devices with this schematic. So, that leaves it out for driving a 2 ohm load of any sort, with the possible exception of a very low power speaker - not an ESL."

Why not? The laterals have very low capacitances. The +-14mA available from the VAS is enough to slew the output at about 100V/uS. (Cgs is mostly bootstrapped away and Cgd is very low) The VAS transistors might actually be what limits the slew rate in this example.
 
megajocke said:
"It would be nice to enforce or create true balance and differential operation of the long tail pairs as a starter point. The lack of a decent Vbe multiplier at minimum for bias setting is pretty outrageous this being 2008..."

Laterals don't need a vbe multiplier. At 100mA Id the temperature coefficient of the current is zero. The bias will vary with + rail voltage though as LTP tail current sets the bias current through the VAS and bias pot too. A vbe multiplier or current source for the diff pair would make the bias more stable with rail voltage.


Exactly.
Stability is something to strive for?

Not having gate resistors on the FETs looks a bit scary though, it might oscillate if the lauout is not perfect.

For protection: Add 10V zeners in series with diodes to limit gate voltage and you will have a very hard time destroying those FETs.

I tried to destroy a lateral fet by mounting it on a very undersized heatsink, connecting 50V drain to source and apply 10V current limited to the gate. The transistor did not blow up - when it got overheated some kind of internal thyristor structure shorts gate-source turning the transistor off until gate current is removed.

Not sure ur english isn't causing a mis-translation here.
Within reason these Mosfets have a negative tempco.
No thyristor I am aware of inside it... ?
Maybe you meant thermistor??
If so, then you mean negative temperature coefficient. :)

One pair driving 2 ohms won't give much power though before current limiting, but if this just is the impedance at the highest frequencies you will probably be fine unless you like high power 20kHz sinewaves :p

Except for one important detail - clipping of high frequencies on an ESL sounds rather awful. Doesn't matter much if it is voltage or current clipping, imho... you'll hear hard clipping, and hear the increase of distortion with level without much difficulty.

"Clearly one is not going to be able to increase the number of output devices with this schematic. So, that leaves it out for driving a 2 ohm load of any sort, with the possible exception of a very low power speaker - not an ESL."

Why not? The laterals have very low capacitances. The +-14mA available from the VAS is enough to slew the output at about 100V/uS. (Cgs is mostly bootstrapped away and Cgd is very low) The VAS transistors might actually be what limits the slew rate in this example.

Ummm?
Not no capacitance though. :D
You scared me, and I pulled my Hitachi book just to check.
The devices called for on the K side are the 2SK1058, spec'd at 600-900pf per gate. (usually around 800pf on a good day)

I guess it depends on how slow you want your amp to go maybe?
I doubt that the amp as shown will slew anywhere near 100v/us.
More like 10v/us.
Dunno.
The usual is a coupla watts or more to shove a row of those around... but ymmv. I don't have the spec's handy on the VAS xistors... my guess was that they were smallish...

Anyhow they show a 3 amp fuse on the rails to the outputs... so it ain't makin much power...

_-_-bear
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.