| Peter Daniel |
It seems that there is certain interest in my work;) , so I present here my actually first creations done some 20 years ago.
My first all out project was a preamp with all possible features. At that time it was running on 220V and was using DIN connectors. You might not believe me but only knobs and a groove across the front face were custom machined, all other openings were drilled out and filed by hand. As I remember it took me a whole day to do it.;) |
|
|
| Peter Daniel |
| All the boards were designed and etched by me and as usually I wanted to avoid excessive wiring. The circuits are arranged in two levels, where bottom contains phono EQ and power supply and top actual preamp and elaborate filters. I think there was not a single feature I didn't put there.;) But then, I was more influenced by Japanese makers then by Mark Levinson. |
|
|
| Peter Daniel |
| At that time I was also into mini systems (315mm width) so both preamp and amp were of that size. The amp was low distortion class B, nothing particular about it if not the small size. It delivered 50 W and I never needed more. The chassis is not done from 1/4" aluminum, but from steel I used from other equipment chassis and cut to size. |
|
|
| Peter Daniel |
| The one thing I like about it, is the compact arrangement and good use of space. Also symetrical layout and separate transformers for ea. channel. If you read the small print on the caps you will know where I'm coming from.:) |
|
|
| Duo |
| Cool stuff, you got a good touch with construction;) |
|
|
| jgwinner |
Very nice!
The front panel looks very nice as well. Rub on transfers for the letters?
Just as a matter of interest, why the different transformers per channel?
One transformer of twice the size is usually cheaper than two, so I'm interested in your opinion given you used two transformers here.
Also, what transformers are those? |
|
|
| Peter Daniel |
The transformers were changed later for 110V since original used 220V. Two of them fitted better in available space, I also think that using separate transformers for ea. channel is better (you know true mono construction). Now I would use much bigger transformers thou.
Dry transfer lettering always worked good for me, and even after 20 years you don't see much wear. |
|
|
| Duo |
I must agree that seperate channel transformers are a very good idea indeed. This method gives greater channel seperation and much better power figures. I have a Harman/Kardon HK402A
and it uses seperate trannies. It's only rated at 60W a channel into 8 ohms, but I can load it with four or 2 ohms and easily get more power than expected. Thus, this amp uses (harman/kardon watts) which is like RMS compared to Peak. |
|
|
| jgwinner |
I'm not arguing, just asking some questions.
If the amp design has a high PSRR, why would separate transformers give bettern separation?
If a channel needs more power, wouldn't be easier to pull from a supply that's twice as large? (i.e. more transient response possible).
What do you mean by 'better power figures'?
I measured my Technics 55w power amp supplies a while ago and was surprised to find they are much higher voltage than expected (I can't remember the number now). It's sized to be a 150w amp as far as the voltage goes, but they rate it at 55w. I'm wondering if this is the same thing you're seeing in the Harmon Kardon. (Not trying to compare the two, it's just interesting).
Part of the reason I'm asking is that for my home theater, 10 power transformers is a whole lot more expensive than 1 or 2 :-) |
|
|
| Duo |
OKay, to answer your questions:
first of all, when you draw large transients of power from a PSU, you'll cause a voltage drop no matter what even if small. WHen you have a seperate PSU for each amp, the amps haven't got much ease in interfering with eachother. Essentially, one amplifier causes a voltage drop on a bass transient or something whil the other is trying to produce vocals, the vocals on the other amp get reduced in amplitude at each transient and this can become noticeable. Basically, it's an easy way to get away with minimal size supplies and still give each amp less disturbance in the line.
For home theater it really depends on what your willing to pay for or listen to. You can achieve exceptional results with only one transformer for a pair or more of amps, but this transformer and it's supply capacitors mustn't let transient voltage drops occur.
Most dolby and DTS recordings have very very bad distortion figures on transients anyway so it doesn't matter too much with home theater. And try not to believe in THX labels either, it's a publicity stunt.
BTW, my next set of amps will run off of the same PSU and trannie and deliver: 2X600W into 4ohms, 2X100W into 8ohms, and 1X80W into 16ohms all from a 5kva transformer but I'm going to use 1 farad supply caps to eliminate transient problems. As you may see, it isn't a necessity to have seperate PSU's, it just a luxury if you can afford it...;)
Well... many limiting factors do occur in amplifiers and also, there are many ways of rating output power. In the Harman/Kardon industry, they often made huge understatements in the ways of power. Also, Harman/Kardon amplifiers are very very well designed for high current capability and excellent current handling. They will not drop off or flake out like many amplifiers these days when you demand bass from them. Thus, they seem way more powerful than the usual amp of the same rating. One good example of HK current abilities is that I could run my amp into a set of 4ohm and 8ohm speakers in parallel which results in a total load impedance of 2.6 ohms. I could crank the amp until it's gain stage distorted and the output transistors didn't even get hot, while if you did this to a modern home reciever or amp, it would turn into a nice fireworks show! |
|
|
| Peter Daniel |
| It all depends on what you whant to get out of your system. I believe that separate power transformers are giving you better value for the money than $300 power cords.;) With them come separate bridges for positive and negative rails (if not discreet soft recovery diodes). At least go for separate transformers for main channels and another one for center and surround. Depending on the rest of your system, but for me its not a law of diminishing returns anymore.;) |
|
|
| capslock |
A large transformer will be more power-efficient and stiffer (i.e. less load dependent) than two smaller transformers of the same total rating.
Hence it might be a good idea both environmentally and sonically to use only on. Depending on how far you want to go, you can always use separate common mode chockes, rectifiers and filter caps for the channels.
In my amp, I have a common rectifier, choke and huge PI-filter. The channel boards are supplied from braided cables (to minimize the magnetic field) passing signal lines only at right angles and input stages only at a great distance. Then each board has its own 3x 4700 uF per rail as local decoupling.
Eric |
|
|
| gromanswe |
Capslock.
I also prefer 1 trafo. That doesn't mean you can't
use separate rectifiers.
One could even provide separate rectifier
for the the inputstage.
To have inputstage well protected from output stage
I find very important.
Of cource different variations could lead to satisfying
result.
Amplified capacitance I also find as a good thing. |
|
|
| Peter Daniel |
Maybe I would use one big transformer for stereo amp now. It was not possible in the pictured amp because my intention was to have very low chassis and two transformers fitted there perfectly. Sometimes looks are important.;)
Today, I wouldn't build a stereo amp. I always go for monoblocks, because I want to keep my amps close to speakers (especially when using expensive cables, when every foot counts).;) |
|
|
| promitheus |
| They look a lot better then my first attempts LOL :-) |
|
|
| gromanswe |
| quote: | Originally posted by HPotter
Today, I wouldn't build a stereo amp. I always go for monoblocks, because I want to keep my amps close to speakers (especially when using expensive cables, when every foot counts).;) |
That is the most sensible thing you can do. It makes you most free to use you endstage in different ways.
In an non-active implication it can different.
Preamplifiers, controlamplifiers there you would like to have
everything in one place.
Not to have to "run around in room" when
selecting different features.
But this could also be overcome by using
remote signals. |
|
|
| gromanswe |
When transfering the signal from signal source,
CDplayer, turntable, to Loudspeaker,
it preferable to maintain the amplitude.
That leaves best margin to background noise.
Background noise can be seen as "the ocean of noise".
It is often expressed in S/N-ratio. Signal to Noise.
When using some meters of Loudspeaker cable,
you get one problem. Transfering high currents
at high speeds.
When using bi-/tri-amping with monoblock mounted
next to woofers, you get another problem.
Tranfering the signal attenuated of Volume control
in centerplaced Preamp, to the Endstage.
My solution is to have a remotely controlled
attenuator, placed next to endstage input.
It can be wire-remote or Infrared-remote.
This solution keeps signal/noise HIGH over
the transfering distance.
The shorter distance a signal is low, the better! |
|
|
| janneman |
Hi Gromanswe,
That is interesting. Can you give us some more info on how you implemented the "local" (to the monoblock power amps) remote volume control?
Jan Didden |
|
|
| gromanswe |
CMOS 40193 up/down counter is remotely controlled
CMOS 4051 8-channel analog multiplexer
resistor divider
1 resistor is in series with incoming signal
where only the 8 resistors are switched to GND
the CMOS FETS works best around GND
this is repeated
twice with a buffer OP in between
to get 8x8=64 steps -1dB/step |
|
|
| janneman |
| quote: | Originally posted by gromanswe
CMOS 40193 up/down counter is remotely controlled
CMOS 4051 8-channel analog multiplexer
resistor divider
1 resistor is in series with incoming signal
where only the 8 resistors are switched to GND
the CMOS FETS works best around GND
this is repeated
twice with a buffer OP in between
to get 8x8=64 steps -1dB/step |
Hmm yes that could work. Are you using R/2R ladder networks? What is the value of your series resistor? What is the ratio of your series resistor to the lowest value switched resistor?
Jan Didden |
|
|
| schizeckinosy |
| quote: | Originally posted by gromanswe
When using some meters of Loudspeaker cable,
you get one problem. Transfering high currents
at high speeds.
When using bi-/tri-amping with monoblock mounted
next to woofers, you get another problem.
Tranfering the signal attenuated of Volume control
in centerplaced Preamp, to the Endstage. |
Hi,
I am interested in this tradeoff, so I started a new thread about it (long cables...) in this forum.
Cheers! Dan |
|
|
| Peter Daniel |
I think that remotely controlled volume control module, right at the input of the amplifier would be the best product ever invented for audio. Lets say you are using tri-amping and you have 6 monoblocks amps. The module would be quite small (maybe the size of a computer mouse or smaller ) and it mounts directly to the input jack of the amp, the interconnect cable is connected then to the module. Then you have a remotely controlled central control console which through the remote, communicates with the modules and sets different volume levels and remembers individual adjustments as well. If you take care of source switching somwhere else you have the best passive preamp ever done. I was playing with that idea for last 10 years, somehow never realised it, maybe because I'm not a microcontroller guy. ;)
I've also never seen it implemented. I'm glad someone else is talking about it too. |
|
|
| gromanswe |
No, 4051 only closes 1 of 8 Jfets-switch.
So only 1 resistor at the time is connected to GND.
Yes, the lowest resistor is the crushial point
when using JFETS with R-on in the magnitude 50 ohms.
To achieve a result satisfactory to your
acceptance of evnetually distortion
you have to calculate how big is R-on
variations in relation to R1+R2+Ron.
A way around that, for ultimate solution is using
DRY reed relays to switch.
This is my favorite
DSS41A05= needs some 8mA to switch
http://www.alliedelec.com/catalog/c...es/2002/712.pdf |
|
|
| gromanswe |
No, 4051 only closes 1 of 8 Jfets-switch.
So only 1 resistor at the time is connected to GND.
Making a conventional 2-resistor divider.
Yes, the lowest resistor is the crushial point
when using JFETS with R-on in the magnitude 50 ohms.
To achieve a result satisfactory to your
acceptance of distortion,
you have to calculate how big is R-on
variations in relation to R1+R2+Ron.
A way around that, for better solution is using
Dry Reed Relays to switch.
This is my favourite
DSS41A05= needs some 8mA to switch
http://www.alliedelec.com/catalog/c...es/2002/712.pdf |
|
|
| janneman |
Gromanswe,
So I probably misunderstood you, I thought you had it working rather than just as an idea. I was interested in your particular choices.
As to the relays, I have posted earlier my preference for a particular type (Pickering) along with the arguments why I have this preference.
Cheers, Jan Didden |
|
|
| gromanswe |
| I have it working. |
|
|
| Variac |
Wow - the best product ever invented for audio!
I guess you need to post a schematic soon.
Of course you still have the problem of the poor little low level signal from the source having to travel to the various locations.
Unless you can split up a cd player! |
|
|
| gromanswe |
I do not claim this to "the best product" for you.
This is what I built over 2 years ago.
A friend who knows how to use CMOS logic helped me
with that part.
It is the best product I have ever built.
That I can claim. |
|
|
| Variac |
| Actually, I was serious- I agree with HPotter- this is pretty cool! |
|
|
| janneman |
| quote: | Originally posted by gromanswe
I have it working. |
Great! Let us have the resistance values for the series resistor and the various switches.
Cheers, Jan Didden |
|
|
| janneman |
Hei gromanswe,
You forgot about your local volume control? Or is it so secret that you want to keep it for yourself?
{Janneman turns to the audience and can be heard muttering "I think he is lying, he has nothing of the sort working. Makes one wonder what else he is lying about". Curtains for Gromanswe}
Cheers, Jan Didden |
|
|
| TomJ |
Peter Daniel,
Can you post a close up of the front of the amp I am interested in what all the knobs and switches are for looks Very cool:cool: :cool: |
|
|
| Peter Daniel |
| One button fell off, that's why the sticker;) |
|
|
| Peter Daniel |
| Here's a better view. |
|
|
| Elkaid |
Peter : how have you written text on aluminium ?
Silkscreen ?
Regards, |
|
|
| Peter Daniel |
| In those days (20 years ago) silkscreening was almost impossible for small DIY like me. So I used dry transfer sheets, available from graphic stores. The brand I used is called Letraset, if I remeber correctly. I'm still using it, as it's fast, easy and convenient method. If handled properly, it's also very durable, as the picture made today shows.;) |
|
|
|
