Roender's FC-100 prototype and builder's thread

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Original the FC100 uses a 2x25Vac toroid is used after the rectifier is this 2x35Vdc.The shunt regulator gives 4 volts more around 39Vdc.I,m planning to us a 2x27Vac transformer after rectifying is this 38Vdc.The front-end is using 4 volts more,my question what to I've to change of the shunt-regulator to get 42Vdc?
 
Patrick, you need not change anything.

Go and use your 27VAC transformer for the backend and the 30VAC that I sent you for the frontend.

I am absolutely sure that Mihai (I consider him to be a real expert) took account of 10% deviation on the primary voltage!

Best regards - Rudi_Ratlos
 
Rudi, I will try it but the frontend transformer is 2x35Vac which gives 49Vdc after rectifying.Is this a problem?
35Vac for the output stage of the amplifier is too high for a 4ohms rated amplifier. The temperature de-rated SOAR gets exceeded too often by reactive IV when working with near maximum volumes.

However, it is my opinion that 35Vac for the whole amplifier would be really good for an 8ohms specified amplifier. If you do not have any 4 to 8ohms speakers and never intend testing any low impedance speakers then building an 8ohms only based on a 35Vac transformer may well give you a very good combination of high sound quality and the ability to send unclipped signal when the going gets towards high SPL.

Conversely, having studied the details, I reckon that a 28Vac transformer would give a very good 4ohms rated amplifier, capable of driving any true 4ohms rated speaker and all the others of higher impedance.
If I had some 4ohms speakers I would be very tempted to use a 30Vac transformer and a very big heatsink, along with 35Vac regulated down for the front end.
 
Last edited:
The front end does not NEED to be supplied with a higher voltage. The amp works well with the same PSU supplying both ends of the amplifier.

One advantage of using a lower front end voltage is that the output stage does not approach rail voltage and thus the output transistors do not approach saturation. When the output becomes very close to rail voltage the transistor betas become lower, the transistors become slower, the capacitance becomes higher. and most importantly if the output clips the stage can become very slow to recover from that clip.

Limiting the front end maximum output such that the output stage never approaches clipping makes for a better "sounding" performance if you like listening at high powers.
 
Last edited:
................Limiting the front end maximum output such that the output stage never approaches clipping makes for a better "sounding" performance if you like listening at high powers.
I should add that listening at low powers never allows the amplifier to approach clipping (both current and voltage clipping) and thus the amplifier always performs as designed/intended.

I wish I could remember who said:
"All amplifiers that do not clip and have the same frequency response sound the same."
or words to that effect.
 
Gentlemen, I am currently performing a group-buy for the FC-100 in the German ANALOG-forum, and I am discussing with the German DIYers,
how a sufficient, low-cost DIY-case for the FC-100 can be built.

We are currently settling on this solution:


http://www.abload.de/img/image11ksar.png

Buy two ALU-plates, serving as side-panels, dimension: about 380x140mm, 8-10mm thick.
Connect heatsinks, whatever is available and adequate, for example this one, cut in the middle, resulting in 200x125x40mm:


http://www.abload.de/img/kk24sjw.jpg

to the "hot-spot" (where the NJL-transistors are mounted) on the outer side of an ALU-side panel and the FC-100 PCB to its inner side.

I think that this solution will do it - and is low-cost as well.

Best regards - Rudi_Ratlos
 
use a thinner side panel, 2mm to 3mm would be sufficient.
Bolt the heatsinks to this side panel.
Mark out where the output devices will sit.

Remove the side panels and cut holes right through, so that the devices can be mounted in direct contact with the sink. Where the side panels bolt to the sinks will give a bit of extra dissipation to the sink and thus the devices will run a bit cooler.

Adding a plate between the devices and the sink does not always result in better cooling.
I think most arrangements with an intermediate plate will actually be worse than without the plate.

BTW,
you are using remote servers instead of attaching.
 
Last edited:
Andrew: I made the experience that the cooling solution that I suggest, does its job very, very well!

Have a look at my construction (attached picture).
I am playing music very loud, for hours, and the 10mm thick ALU-plate does not even get "hand-warm", as well as the attached heatsinks do not!

I applied a layer of thermal grease between ALU-plate and heatsinks and screwed them very tightly together using 12 bolts (per heatsink).

I therefore believe that my construction even is "thermal overkill".
A 6mm thick ALU-plate with heatsinks attached to the hot-spot (where the NJLs are mounted) will do it as well.

Best regards - Rudi_Ratlos

P.S. Lucylu: go on, please!
 

Attachments

  • Eloxiert 001.jpg
    Eloxiert 001.jpg
    395.9 KB · Views: 857
Last edited:
They look similar to mine.
I have a pair of 250 high by 200 wide by 40mm thick. 20off 30mm fins on a 10mm backplate.
At the full 185mA and supplies at +-42Vdc (quite a bit higher than using a 25Vac transformer) the sinks are too hot to hold longer than about 5 seconds at the cooler edges.
The lower voltage of the standard build and the 160mm high will probably end up at about the same temperature.
The 3pair output stage is very robust and can be very hot and yet still stay within an acceptable, good reliability SOAR.
Don't be too afraid of a sink that is just a bit too hot to hold.

Too hot to touch is a bit hotter.
 
Gentlemen, I have sent you an EMail an hour ago telling you about the mishap (error) in my FC-100 PCB layout.
I have posted a bug-fix ("free-air solder-style") in my EMail.

Another, more professional bug-fix would look like the one in the attached image.
Cut the trace at the position labelled with the pink "x". Drill a new 1mm-hole into the PCB (this PCB is not the MonaLisa !)
directly underneath the cut and solder the lower lead of the MKP1839/1µF into this hole.
There may be even "better bug-fixes" (?).

I once more regret for causing you any inconvenience.

Best regards - Rudi

P.S. Affected by this "mishap" are those of you, who regularly drive the FC-100 into clipping.
 

Attachments

  • Reparatur1.PNG
    Reparatur1.PNG
    60.9 KB · Views: 599
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.