***Update 10-12-2014: Updated stuffing guide***
The silkscreen for the amplifier PCB contains an error where the labeling of L1/R6 is swapped, this is corrected in the attached stuffing guide with the _001.pdf revision number. Please use this for stuffing and disregard the previous version. New version is in post #510. If you've already built the amplifier, please swap L1/R6 around as the amplifier will see a large(r) DC offset and reduced performance due to this error.
http://www.diyaudio.com/forums/solid-state/96853-extrema-class-strikes-back-51.html#post4147769
***Update 01-12-2014: parts value change***
With a simple parts value change the ExtremA amplifier sees an improvement in slew-rate, step response and overall stability, so it is recommended making these changes. There's four capacitors that need to be changed:
C1/C2 = 68pF - COG or NPO - 5mm pitch (Vishay NP0 ceramic)
C4/C9 = 220pF - film capacitor, polypropylene or polyester - 5mm pitch (WIMA FKP2)
***
Alright, I nicked that subject title from someone else, kind of cheesy perhaps, but I like it.
For those of you that can appreciate a properly designed class-A amplifier with a spec sheet and performance figures that the vast majority of commercial amplifiers will never be able to match, take a look at the below url. And yes, to answer the inevitable question that's bound to pop up, there's PCBs available.
But before you get all excited, be warned, this is NOT a project for the novice builder.
ExtremA, a reference class-A DIY amplifier
ExtremA, a reference class-A DIY amplifier. - Introduction
Best regards,
Sander Sassen.
The silkscreen for the amplifier PCB contains an error where the labeling of L1/R6 is swapped, this is corrected in the attached stuffing guide with the _001.pdf revision number. Please use this for stuffing and disregard the previous version. New version is in post #510. If you've already built the amplifier, please swap L1/R6 around as the amplifier will see a large(r) DC offset and reduced performance due to this error.
http://www.diyaudio.com/forums/solid-state/96853-extrema-class-strikes-back-51.html#post4147769
***Update 01-12-2014: parts value change***
With a simple parts value change the ExtremA amplifier sees an improvement in slew-rate, step response and overall stability, so it is recommended making these changes. There's four capacitors that need to be changed:
C1/C2 = 68pF - COG or NPO - 5mm pitch (Vishay NP0 ceramic)
C4/C9 = 220pF - film capacitor, polypropylene or polyester - 5mm pitch (WIMA FKP2)
***
Alright, I nicked that subject title from someone else, kind of cheesy perhaps, but I like it.
For those of you that can appreciate a properly designed class-A amplifier with a spec sheet and performance figures that the vast majority of commercial amplifiers will never be able to match, take a look at the below url. And yes, to answer the inevitable question that's bound to pop up, there's PCBs available.
But before you get all excited, be warned, this is NOT a project for the novice builder.
ExtremA, a reference class-A DIY amplifier
ExtremA, a reference class-A DIY amplifier. - Introduction
Best regards,
Sander Sassen.
Last edited:
SSassen said:
It does not look too bad.
No even to our Lineup Audio Lab
Class A amplifiers.
you have done a very good job, by most standards
not at least the magnificent web presentation
Regards
lineup
Leolabs said:
Refering to this schematic:
http://media.hardwareanalysis.com/articles/large/11997.gif
There are a lot!!!! of small signal transistors.
What pair of transistor do you mean, Leolabs?
Hi Sander,
From your experience with this design, would you advise a similar "bridged" output topology for a very low < 0.5 ohm speaker load? The extra power dissipation from the lower efficiency compared to a traditional grounded-output topology is the concern. I'm currently using a symmetric JFET-Bipolar folded cascode with a bridged 16 Sanken Bipolars output topology and really big heatsinks to put 32 Class-A watts into 0.15 ohms. The output stages run from +/- 5V power supply, so each diode drop is power expensive.
FROM ARTICLE: "Neither input nor output is ground referenced. This takes the influence of ground currents and loops out of the equation. The only thing that is referenced to ground is the common mode voltage of both outputs, this is done to insure that both outputs have the same dynamic range and the output remains symmetrical. Since the load is attached between the two outputs, it is not interested in the common mode voltage."
From your experience with this design, would you advise a similar "bridged" output topology for a very low < 0.5 ohm speaker load? The extra power dissipation from the lower efficiency compared to a traditional grounded-output topology is the concern. I'm currently using a symmetric JFET-Bipolar folded cascode with a bridged 16 Sanken Bipolars output topology and really big heatsinks to put 32 Class-A watts into 0.15 ohms. The output stages run from +/- 5V power supply, so each diode drop is power expensive.
FROM ARTICLE: "Neither input nor output is ground referenced. This takes the influence of ground currents and loops out of the equation. The only thing that is referenced to ground is the common mode voltage of both outputs, this is done to insure that both outputs have the same dynamic range and the output remains symmetrical. Since the load is attached between the two outputs, it is not interested in the common mode voltage."
"Neither input nor output is ground referenced. This takes the influence of ground currents and loops out of the equation.”
Not entirely. Unless driven from a differential-output preamp, one of the inputs will need to be tied to ground, thus referencing the other input to ground.
Cheers,
Glen
Not entirely. Unless driven from a differential-output preamp, one of the inputs will need to be tied to ground, thus referencing the other input to ground.
Cheers,
Glen
Oh my, you must've fluked math class at school, let me calculate the price that's associated with this project for you:
The article notes the total price of a mono amplifier to be about $700 (USD) or 500-euro, this includes the PCBs. The PCBs themselves are double sided, silver immersion over 35u copper, with metallization, solder mask and silk screen.
A popular PCB manufacturing outlet that also does small runs for hobbyists is http://www.thepcbshop.com, they're generally considered as having very affordable pricing. Suppose you want two sets of PCBs made, as you're building two amplifiers. According to their pricing, you'll be paying:
- 2x double sided PCB, 105mm x 135mm (amplifier), 15 day delivery, their cheapest option: 69.93-euro
- 2x double sided PCB, 135mm x 155mm (powersupply), 15-day delivery, their cheapest option: 81.56-euro.
Total of 151.49-euro
Hence our offer certainly isn't more expensive. Granted, we could get a discount if we ordered more PCBs, 25, 50, or 100-pcs does bring the price down. But since we ordered just a couple to be able to supply interested DIY-ers with a set of PCBs we're just earning a few euro's per set of PCBs, as that price includes shipping worldwide.
If we sell a lot of PCBs, we might be able to break even with the development cost of the prototypes. However, the time invested in the creation of this design presented for free to the DIY-community we'll never get reimbursed, or at least not at our normal hourly rate. So we'll just write that off as being time spent on our hobby I guess.
Thanks for appreciating the fact that we're offering this design free of charge, some people tend to think that everything should come for free. So my question to you would be, what have you contributed to this community lately?
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
LineSource,
Well, you'd need to calculate the requirements for bias current and power stage supply voltage as per the example in the article, see what that gets you and what the net efficiency would be. Low impedances are always a bit of challenge, Douglas Self has a number of pages dedicated to the topic in his book, and rightfully so.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Well, you'd need to calculate the requirements for bias current and power stage supply voltage as per the example in the article, see what that gets you and what the net efficiency would be. Low impedances are always a bit of challenge, Douglas Self has a number of pages dedicated to the topic in his book, and rightfully so.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Glen,
Well yes, if you use this amplifier in a single ended configuration. It however is designed for and targeted at the far less problematic balanced operation as is common in studios and with more expensive equipment. As really, whoever came up with the idea of using GND as a signal reference and a means of equalizing chassis potential probably had an off-day at the office.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Well yes, if you use this amplifier in a single ended configuration. It however is designed for and targeted at the far less problematic balanced operation as is common in studios and with more expensive equipment. As really, whoever came up with the idea of using GND as a signal reference and a means of equalizing chassis potential probably had an off-day at the office.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Leolabs,
As mentioned in the article, only the input pair, which are BC550C transistors. You could go through the trouble of matching all of them, but it will yield just a small benefit as per the comments about the distortion plots. Mostly you'll see a reduction in initial offset, but that's compensated for by setting the DC offset trimpot, hence a non-issue.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
As mentioned in the article, only the input pair, which are BC550C transistors. You could go through the trouble of matching all of them, but it will yield just a small benefit as per the comments about the distortion plots. Mostly you'll see a reduction in initial offset, but that's compensated for by setting the DC offset trimpot, hence a non-issue.
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Quote :
"this is NOT a project for the novice builder"
http://www.diyaudio.com/forums/showthread.php?postid=1142106#post1142106
Quote :
"the most Hi-End project"
http://www.diyaudio.com/forums/showthread.php?postid=1142668#post1142668
Sorry for my poor English, Meneer Sassen, but I do not consider the two description as listed above contradictory.
And I couldn't quite find the word costly or expensive anywhere, could you ?
> you must've fluked math class at school
Oh yes I did, but I still somehow managed straight A's for A Level, 1st honour at University, Ph.D. in Precision Engineering, some 40 patents worldwide. Just lucky I guess. 🙂
> So my question to you would be, what have you contributed to this community lately?
Not too long ago, a harmless little circuit.
PCB layout also published, just by pure chance.
http://www.diyaudio.com/forums/showthread.php?postid=1133142#post1133142
Patrick
"this is NOT a project for the novice builder"
http://www.diyaudio.com/forums/showthread.php?postid=1142106#post1142106
Quote :
"the most Hi-End project"
http://www.diyaudio.com/forums/showthread.php?postid=1142668#post1142668
Sorry for my poor English, Meneer Sassen, but I do not consider the two description as listed above contradictory.
And I couldn't quite find the word costly or expensive anywhere, could you ?
> you must've fluked math class at school
Oh yes I did, but I still somehow managed straight A's for A Level, 1st honour at University, Ph.D. in Precision Engineering, some 40 patents worldwide. Just lucky I guess. 🙂
> So my question to you would be, what have you contributed to this community lately?
Not too long ago, a harmless little circuit.
PCB layout also published, just by pure chance.
http://www.diyaudio.com/forums/showthread.php?postid=1133142#post1133142
Patrick
Patrick,
I'm sorry, you lost me, what are you referring to exactly? 😕
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
I'm sorry, you lost me, what are you referring to exactly? 😕
Best regards,
Sander Sassen
http://www.hardwareanalysis.com
Hi,
Read through ur article and looked at the schematic... very interesting and nice design. The results (graphs) look quite impressive!
Was wondering, have you measured the slew rate, how about the power bandwidth?
Also, wondering what a square wave looks like on a fast scope with a highish speed squarewave input?
How about stability into relatively capacitive loads?
In terms of design, to what do you attribute the rather low distortion figures? I see extensive use of current mirrors throughout the input stages... and asking from a position of admitted ignorance here, is this design idea used elsewhere (like an opamp circuit?)
Oh yeah, what's the DF look like?
No benefit to paralleled outpoot devices in your view?
Do we get to see a jpeg of a real world implementation?? 😀
Nice work!
_-_-bear
Read through ur article and looked at the schematic... very interesting and nice design. The results (graphs) look quite impressive!
Was wondering, have you measured the slew rate, how about the power bandwidth?
Also, wondering what a square wave looks like on a fast scope with a highish speed squarewave input?
How about stability into relatively capacitive loads?
In terms of design, to what do you attribute the rather low distortion figures? I see extensive use of current mirrors throughout the input stages... and asking from a position of admitted ignorance here, is this design idea used elsewhere (like an opamp circuit?)
Oh yeah, what's the DF look like?
No benefit to paralleled outpoot devices in your view?
Do we get to see a jpeg of a real world implementation?? 😀
Nice work!
_-_-bear
The price is very dependent of how many pcb's you dare to order and of course the desired quality.EUVL said:
1. too many gain stages, too many transistors,too many parts....
2. what`s the output power working in pure class a?
3.Isn`t that heatsing small for class a?
4.one bjt output pair?
5.most hi end? - Don`t get me wrong, I didn`t hear your amp and maybe the sound is great but... "most hi-end" seems funy to me...
cheers!
2. what`s the output power working in pure class a?
3.Isn`t that heatsing small for class a?
4.one bjt output pair?
5.most hi end? - Don`t get me wrong, I didn`t hear your amp and maybe the sound is great but... "most hi-end" seems funy to me...
cheers!
... but the important thing is the end result, not how to reach it. I have great confidence in Bruno Putzeys. I'm sure that he has put some thought into the design. I'm also sure it's possible to reduce the amount of parts but this isn't a commercial research project. It's a crazy idea made real.bogdan_borko said:
SSassen said:
OK, but you are promoting the design here on a DIY board. I'd say most people here drive their HiFi amplifiers from non-balanced signal sources. That goes for those who have built themselves mega-buck amplifiers too.
BTW, I'm not trying the criticise the design. Driven either way, as far as I can see, it will still work very well indeed.
Cheers,
Glen
- Status
- Not open for further replies.