Found this old Elektor amp

Edit:
I don't want to build this amp, I did a long time ago. I just found it in my pc and was curious about if people built it or ideas on it.

Hi, I found this in my pc as I was cleaning a bit.
I remember I built it a long time ago.
At first glance I tought the intput is a diamond buffer, but it's just a push-pull differential pair ? T7 / T8 VAS .
T5 / T6 being current sources for the input pairs ?. I don't get what the 5v6 zeners do tho.
( I usually used Constant current sources with two transistors like this:
CCS.JPG


D3/D4/D5/D6 limit the output current?.
How did they get away with only 3x820nF C3/C4/C5 ( 2.4uF ) and non polar in the feedback , I usually see polarised in the range of 100-200uF there.
I know it's an old design, from Elektor as I recall. using the nostalgic 2n3055's .

- Bruno.
 

Attachments

  • 100 watt amplifier circuit.jpg
    100 watt amplifier circuit.jpg
    324 KB · Views: 964
Last edited:
For the CCS, 5V on 10K is 0.5mA as the schematic shows, but 0.65V on 260 is 2.5mA so I assume you used something more like 0.65/0.0005 = 1300.
4k7 and 2.4uF is about 14Hz. Normally R4 is smaller so the decoupling cap has to be larger. On a single sided LTP, there is a slight DC bias which makes using a polarized cap OK, although diode clamps are a good idea to prevent over-voltage on the LTP(s).
R13 and R14 are a bad idea that makes an unnecessary pole in the feedback loop.
Pot P1 failure mode will cause output overcurrent and likely blow up. Any VBE multiplier would be a better idea, which could be a symmetric if you like, but I would recommend a CFP VBE multiplier for best bias regulation. This is especially important because the VAS current is determined by the LTP current, ie the classic problem with a symmetric IPS. Your use of a transistor CCS is actually a source of some thermal compensation, vs the Zener CCS.
Yes the diode drops of D3,4,5,6 plus (half) the bias voltage are a voltage limit on the 1/3=0.33 Ohm emitter resistance, hence a current limit, about (3x0.65/0.33=) 5.85 Amps, but I would be surprised if it was actually close to that figure, and it will be temperature sensitive. The clever placement of these diodes means that shoot through current limit is half the output current limit, but a bias (P1) failure negates the current limit. A spice simulation would be wise to verify these things.
 
Last edited:
@steveu is this the way to use the speed up cap?. (pls click to enlarge)
hh.JPG
In this config if the Pot goes open, the output current will colapse. but it's weird that the lower the resistance the higher the current instead of the higher the resistance the higher the current. I put R8 to kind of limit the current to 500mA if the pot goes short circuit ( i don;'t think it can?, they go open most of the time right?). Anyway, I'm building a simple amp for a small sub, in the 70-100W range, but I want it to be stable , thermally and fool proof. if you can elaborate on that VBE multiplier , good practice to keep it separate or on the same heatsink with the drives, but not on the output transistors, all transistors on the damn heatsink. what would improve thermal stability, not prone to thermal runaway, but also not reduce the quiss current so it sounds like crap if the amp gets hot, is a way? or compromises. I haven't really read much about biasing, diodes, vbe multipliers and how they " tame" the outputs not to thermal runaway, blow, etc.
 
Have a read if this thread:
https://www.diyaudio.com/community/threads/two-transistor-vbe-multiplier.388578/
If you have a large heat sink, it becomes less critical. If it's too small, nothing will help. Simulation can predict the bias drift by "step"ping the "temp" variable, but that may assume an even temperature distribution. No substitute for testing in an oven/chamber.
If you are making a sub, consider class-D where it's less of an issue.
 
  • Thank You
Reactions: xXBrunoXx
Hi, maybe my experience with ELEKTOR will make these problems more understandable for you.
Elektor was one of my major information sources many years ago (think from 1975 on). I collected any edition for 25 years or so.
Me and some friends build many of their audio projects and allmost always we had problems with them functioning reliably or at all. Oscillation of amps was considered something normal, you had to fix it your self. Osciloscopes were rare and expensive at that time, so many did'nt ever notice and simply bought new tweeters on a regular base.
In most cases further editions of the magazine had corrections in them. They usually started: "Probably to differences in quality of electronic parts, some readers building this project experienced some minor problems...". Followed by some bug fixes that sometimes cured problems, in most cases not.
Even as this was frustrating, it pushed me to building electronics with highest precission and care for the last detail. So my build quality could not be the cause for missfunction. I once had an professionel electronics developer visiting me, who did not believe me to have soldered my PCB`s by hand.

Sometimes, when an Elekor project didn't work, I was so brazen to make telephone calls to the developers, which was possible at that time, but expensive due to high communications cost of long distance calls. As my father, a journalist, had an astronomic telephone bill anyway, I didn't care.
That made me understand how ELEKTOR worked: They got projects offered by different developers and hobbyists. When interested in publishing, they got paid no (Hobbyist) or very little money (Developer) to do some Photographs, schematics, PCB etc. and write a text about the build. We talk about very low 3 digit German Marks at that time.
Then projects got published without the ELEKTOR staff building or testing any of these for verifying them working reliable. Only sometimes they got a sample for testing. Audio projects where considered lowest value, digital stuff and measuring equipment got more attention. By the way, many of ELEKTOR's staff own developnets where only done on paper, never build at all as published.
Often important, critical parts the external developer wrote, were edited out to make it fit page size.

It was (and is) all about selling a magazine and PCB's, not highest quality or 100% working products, as there was no way to make them responsible for their faulty advice. They could blame allmost any problem on bad build quality and low quality parts. Then, there was no Internet, so hardly no communication between builders. They got allong with that sheme for decades. Of course, they take no responsability for radiation problems, which are caused by following their plans and can cost you a fortune in EU countries. D-amps anyone?

So if your ELEKTOR build does not work or has obvious design faults, this is not something extraordiare. We did not even have any software at that time, remeber! If you consider building any ELEKTOR stuff, look for an alternative first. Nothing they offer is even comparable to todays refined DIYS stuff we get from diyAudio and people being present here.
 
Last edited:
@Turbowatch2 Also weird, I assumed if a magazine publishes about something they actually test what they publish.
We had in Romania, the magazine " Tehnium " from the 90's untill 2006 or so. it was popular in the 2000's because not everyone had internet , and also most of the schematics, were working and made by proffesors, but yea , also projects, schematics submitted by readers I think. I did not get to be in that time , I'm younger (27) But I remember I had a colection of them in a box from my dad.
 
Hi, maybe my experience with ELEKTOR will make these problems more understandable for you.
Elektor was one of my major information sources many years ago (think from 1975 on). I collected any edition for 25 years or so.
Me and some friends build many of their audio projects and allmost always we had problems with them functioning reliably or at all. Oscillation of amps was considered something normal, you had to fix it your self. Osciloscopes were rare and expensive at that time, so many did'nt ever notice and simply bought new tweeters on a regular base.
In most cases further editions of the magazine had corrections in them. They usually started: "Probably to differences in quality of electronic parts, some readers building this project experienced some minor problems...". Followed by some bug fixes that sometimes cured problems, in most cases not.
Even as this was frustrating, it pushed me to building electronics with highest precission and care for the last detail. So my build quality could not be the cause for missfunction. I once had an professionel electronics developer visiting me, who did not believe me to have soldered my PCB`s by hand.

Sometimes, when an Elekor project didn't work, I was so brazen to make telephone calls to the developers, which was possible at that time, but expensive due to high communications cost of long distance calls. As my father, a journalist, had an astronomic telephone bill anyway, I didn't care.
That made me understand how ELEKTOR worked: They got projects offered by different developers and hobbyists. When interested in publishing, they got paid no (Hobbyist) or very little money (Developer) to do some Photographs, schematics, PCB etc. and write a text about the build. We talk about very low 3 digit German Marks at that time.
Then projects got published without the ELEKTOR staff building or testing any of these for verifying them working reliable. Only sometimes they got a sample for testing. Audio projects where considered lowest value, digital stuff and measuring equipment got more attention. By the way, many of ELEKTOR's staff own developnets where only done on paper, never build at all as published.
Often important, critical parts the external developer wrote, were edited out to make it fit page size.

It was (and is) all about selling a magazine and PCB's, not highest quality or 100% working products, as there was no way to make them responsible for their faulty advice. They could blame allmost any problem on bad build quality and low quality parts. Then, there was no Internet, so hardly no communication between builders. They got allong with that sheme for decades. Of course, they take no responsability for radiation problems, which are caused by following their plans and can cost you a fortune in EU countries. D-amps anyone?

So if your ELEKTOR build does not work or has obvious design faults, this is not something extraordiare. We did not even have any software at that time, remeber! If you consider building any ELEKTOR stuff, look for an alternative first. Nothing they offer is even comparable to todays refined DIYS stuff we get from diyAudio and people being present here.
This summarizes it nicely. I disgust all this Elector BS. I preferred the German magazine ELRAD (Heise) being a serious alternative at that time.
 
That could be perceived as a detail but FI the value of R8 is 10x lower than what would be necessary, that s just a detail among others that Elektor designs are not that well thought, they described a ton of amplifiers but there s often several flaws in their designs, FI their Crescendo amp was horribly designed and a recipe for oscillations, they gave some technical explanations about their particular compensation, or rather lack of, but actually they didnt master all the holding and borderings of amplifiers design.

That being said this design could work but the VBE multiplier is of poor conception, also the output stage is dubbious, at the time there were complementary devices for the 2N3055 like the MJ2955, dunno why they used a quasi complementary topology, a triple EF using a BC546/BD139/2N3055 and its complementary combination BC556/BD140/MJ2955 would work way better at a lower distorsion and better stability.
 
That top half of the output stage is 99.999% of the problems. It’s not even the proper way of making a CFP driver.

Changing to regular EF3, or even QC with the EF3 on top and the bottom like they have it normally works the first time every time.
 
Elektor (originally called elektuur , a dutch electronics magazine founded by Bob vd Horst) was nicknamed the "error magazine". And you always had to wait for some more months, issues to read the rectifications named Elektor Leaks (Het lek van Elektuur). It was very innovative, I remember having read an article describing a classD amplifier using a 2N2219 and 2N2905 fast switching transistors. this was before 1970. Has anyone that article?
 
thanks, dutch is my native language....
At the time there was an electronics shop called VanDam electronica. John van der Sluis (RIP) used to work there and they sold an amp kit based on Quad 303 schematics. This became later hawk electronics when he joined nico vis, the former repair technician at van Essen , the quad importer in rotterdam.

I built my first 303 clone using my own pcb when I was 13 years old. The darn thing pulled a lot of dc (at the time I did not know it was oscillation). John vd Sluis helped to stabilize the circuit, but eventually he gave me the official pcb for free, and that worked as promised. Myfirst amp was an ecl82 based circuit, I built that one when I was 9.