I love well done jobs! Possible having a picture of the inputs/outputs side?Finally after finishing this amp, the whole story, beware 🙂
Studying and simulating this schematic it occurred to me that this design of AKSA and Nico Ras is --deceptively-- simple. Although it consists of only two BJT transistors and some other components, the value and probably also the quality of all components determine the outcome. So that is the case with the U2/R12 combination, the choice of transistors, the use of R13/C1 and R7, R5/R6, filtering the supply voltage and probably much more that I am not aware of.
View attachment 1361440
I made the harmonics profile changeable by using a multi-turn pot (U2). The combination U2/R12 set to 27k results in nice harmonics and sound using my 80ohm headphones at 12V supply voltage. If you use headphones with a different impendance you may want to change the value of U2 to your liking. Or, if you want the least distortion possible you can set that as well. Increasing the supply voltage also changes the harmonics profile. So there are a few different, simple ways to modify the workings of the circuit, making this a flexible design.
View attachment 1361442
View attachment 1361443
Originally I removed R13 and C1 to save some space on the small 4x6cm prototyping board. I should not have done that. Testing it I noticed signs of oscillation so I had to add R13 and C1 again, and I also added R7 just to be sure. That helped to get rid of any possible oscillation.
The two TIP42C transistors are mounted onto the bottom of the case, sufficiently cooling them. It becomes barely warm over time.
I mounted the R5 and R6 5W resistors vertically to create some space on the board.
L1 and L2 are in fact one ferrite ring, + and - both wind around that ring, 14 turns of any isolated wire that you can find. Internal wires from UTP network cabling work well.
Be sure to connect the signal ground of all 'sound-carrying' components (volume pot, RCA connectors, output connector) to the correct, "clean" side of the ferrite ring, that is the left side of the inductor as seen on the schematic. If you do not you shall likely end up with a motorboating amp. The connection of GND to the aluminium case is switchable.
You may ask yourself what C7 is for. The 1nF capacitor prevents crackling sounds if turning the volume pot. I have the impression not everyone has that problem but I have, so I added it.
Just like the Formula3HP amp of Lineup that I made before I used a cast-aluminium case again to build it like a tank. One may call it industrial design but let's be honest, it is the best I can do to give it a somewhat nice robust finish without having too many tools to accomplish that.
After sanding and brushing the aluminium using WD-40 as a lubricant I coated it with, you guessed it, WD-40.
The color of the aluminium may become a bit darker over time, which I like because it gives some character to the design. After some time you can apply WD-40 again but I noticed that is not needed if the amplifier is not handled too much.
The enclosure is a Velleman G120B: 'sealed die-cast enclosure - aluminium - 171 x 121 x 55 mm', see https://www.velleman.eu/products/view?id=7327&lang=en. It is not the easiest case to use because the inside is not regular with all the protrusions and it is small. It was a bit of a puzzle to fit everything in but I somehow managed to do that.
View attachment 1361445
View attachment 1361446
View attachment 1361448
In the bottom of the enclosure I made a 1cm hole beneath every board. Four 0.5cm holes at the top of the case then hopefully create a 'chimney effect' to cool the inside adequately. The holes in the prototyping boards that are left unsoldered help to distribute the warmth.
All this may not be necessary using 12V supply voltage but at 19V to 24V I expect it shall.
View attachment 1361447
The boards are not fixed but fastened enough not to become loose. Especially the amplifier board is able to move vertically ever so slightly as the temperature of the transistors change. Silicone tubing is used as board standoffs.
View attachment 1361449
View attachment 1361461
View attachment 1361462
View attachment 1361456
Now the purpose of it all, the sound. First, there is hardly a thump switching it on or off. Then, the first words that come in mind are 'accurate', 'detailed', 'natural', and I should add the word 'very' to that, despite knowing the harmonics profile. Now I know what Michelag means when he calls it the 'Wire'.
I can clearly hear the acoustics and sometimes the sound extends beyond the headphones. If I get a shiver down the spine with a smile on my face there's nothing more to be desired...
Enough written and enough read. All in all it took some thought, effort and time to finish building my version of this amp but it is more than worth it! I am certain this amp deserves its own official name and its own thread.
Thanks again AKSA and Nico Ras!
Ah,and please, a mouser code for the poly switch and ferrite? Thanks!
Thanks. Not much to see at the back, just one pair of RCA input, 6mm jack output, a switch to connect the chassis to earth.I love well done jobs! Possible having a picture of the inputs/outputs side?
Ah,and please, a mouser code for the poly switch and ferrite? Thanks!
For supply power input I used an 3-pins aviation connector GX16 because that is isolated from the chassis.
The poly switch is a RXEF040 type.
All ordered from the far east...
B.t.w. one of the photo's shows the inside/back.
The ferrite ring I got from an old PC power supply. It has a yellow color, possibly an iron powder core, but it seems to work well.
Last edited:
Well spotted Michelag! When I was testing it the circuit was motorboating like crazy and some components became hot. Should not happen, I thought, because I remembered AKSA saying to you 'typically you would use at least 10uF per ma in the supply'.Ah, @widea, wait. I've got two 3300uF caps in the power supply, since with 2x2200uF motorboated. Now I see you're using 1x2200uF for both channel, well, keep an eye on it...
I'm running at 24V.
LTSpice says one circuit draws 93mA at a supply voltage of 12V.
So that needs one capacitor of 2 x 93mA x 10uF = 1860uF. 2200uF should be more than enough using it at 12V supply voltage.
And that was true: after I connected signal ground to the correct, 'clean', side of the inductor, motorboating stopped and everything became normal.
The circuit drew m u c h more current than usual when motorboating. Even 4400uF was not enough.
Yes, this mixed bag of a project of mine is my 'holistic' approach of things 🙂Where did you get those "vintage" Wima output caps ? 😁
I got these vintage Wima capacitors from the only classic electronics shop left here in the city where I live, The Hague, Holland.
They are rated at just 10V but they only have to cope with 5.7V, doable.
These are not the only old/vintage components in the circuit. The 1nF capacitors at the input are very old as well, but still good quality. And the resistors are part of the collection that I have, more than 25 years old and part I got at the before mentioned shop.
I take into account that old components like the Wima capacitors fail over time, perhaps more quickly than recently manufactured ones. That is part of the reason I placed them on a separate board. I test and pair all components before I use them and I give the amplifier a burn-in. Yet I have to see a component fail, but time will tell.
Last edited:
@AKSA asking out of curiosity, would a ferrite bead be just as effective as a base stopper resistor?Sometimes a base stopper is helpful to prevent oscillation. You might use around 47R carbon for this role however.
I have not tried a bead but it would be better at very high frequencies, maybe beyond 1MHz. For audio I would use a carbon resistor, and for RF a ferrite bead.
Hugh
Hugh
Thank you for your answer Hugh. I want to experiment using a D45H11 BJT for Q2 which has a bandwidth of 40 Mhz. Simulation shows a nice second harmonic and hardly anything else (I changed the circuit somewhat to achieve that).
As I understand it, using a faster BJT like a D45H11 increases the chance of oscillation. I will probably use a ferrite bead in series with a 47 ohm carbon resistor at the base of both Q1 and Q2 and see what happens.
Cor
As I understand it, using a faster BJT like a D45H11 increases the chance of oscillation. I will probably use a ferrite bead in series with a 47 ohm carbon resistor at the base of both Q1 and Q2 and see what happens.
Cor
What did you change? Can you post a schematic?Simulation shows a nice second harmonic and hardly anything else (I changed the circuit somewhat to achieve that).
Sure, see A bit of the Hakuin HPA and a bit of the KISS HPAWhat did you change? Can you post a schematic?
- Home
- Amplifiers
- Headphone Systems
- DIY Class A Headphone Amp suggestion