Serial or parallel

[K-amps]

Originally posted by K-amps
Jacco/Andrew/Stuart and others:

(Pinky feel free to delete this post as it does not relate to the KSA50 but then where will I find such talent in one thread. )

I have a couple of Nelson's old Forte 3's. (200wpc -AB) they are running a 2 ohm load is taxing them (Sonically). Using just one channel of each amp... other not loaded.

I cannot help but think this is not efficient use of 2 power amps.

I have 2 options: Parallel bridge 2 channels to make one hefty channel. (By decoupling one channel's input VA stages and running both OP stages by the other channel's input/VA stage.)

Option 2: Run the Trafo primary in series, reducing the rails by 50% and running both channels serial bridged.


Questions:

1) Under which arrangement will I be able to get more class-A biasing (I am guessing parallel)
2) Will the serial bridged OP be able to run a 2 ohm load?
3) What kind of class-A levels can I expect from the serial bridge arrangement.

The current OP stage is 4 pairs of the 3281/1302's running off 75 vdc rails. After serially bridging them there will be 8 pairs running off 37.5 vdc rails.

I am leaning towards bridging because i have always liked the less grainy sound of many bridged amps (as long as the loads were not hairy and the rails were not high).

Once again apologies for sticking this question in this thread. you guys can reply on my personal email.

Best,
K-fed-up

Originally posted by jacco vermeulen
K,

How do you spot a con-artist ?
Easy, wooden shoes don't have laces !

Seriously,
if your Forte Audio M3 didn't hit that thermal braker on the heatsink and those nice Toshies didn't turn Central-Africa then the PS saved them because it's anaemic at 2 ohms use.
Which is not a miracle, i had one of the earlier Class A Forte models.
Worked great in low imp loads, but had about the same layout with way lower rail voltages.
(i used to have dreams about glueing a piece of paper on the front with SA-Slash printed on it)
Those big wattage Forte amps were supposed to run normal loudspeakers, not the crematory.

Whether you half the rail voltages and bridge it, or place the output stages in parallel, makes no difference.
Whether half rail voltage/number of parallel devices or full voltages/number can make a difference for the Tosh SOA.
But you'd still have the 2 Ohm load, and 75V across it.
=> same peak current=> same load the PS can't deliver.

Class A bias level makes little difference, whether parallelled or bridged. Expect a lower clean power figure when bridged, see my response to Stuarts post on bridging.

You can place a bull behind a cow, but if he did not have his cereals there be no flaggin the queen.
If you need more current:
- bigger toroid
- lower the effective voltage across the load

The pic down below is a bridged 100/8 monaural amp that had those Toshiba numbers.
It could be set to 2/1 ohms use through a lower PS voltage, 100Wattsthat in 2 and 160W in 1 Ohm.
1/2 the rail voltage and 2 times the output current=> same VA rating for the transformer.
That musician/designer guy from the Old Reich again

[K-amps]

Well Jacco:

The PSU while Anemic is capable of more (Or I can change the tarfos no sweat).

Issue is: I think I am running the hfe's of the OP stage into non-linearity.

Tying the 2 channels will give me a couple of advantages:

1) Gain of the OP stage will be enough to drive 2 ohms... maybe?
2) Additional heatsinking available means I can bias it to some extent into class-A, making the rails a little more ridgid thereby giving me taut bass.
3) The extra pair of power caps that are loafing around are now in use.

The load is not tru 2 ohms... more like 2.66 ohms resistive with a 0.43 ohm inductor in series. (All resistive).

I am open to changing the trafos ... what i am trying to get at is a PSU and OP stage that is loafing when driving 2 ohms...

[K-amps]

By the way does anyone have the schematics of the Forte model 3?

[jacco vermeulen]

Arif,

At 7 amps output current the old Toshibas still had hFe values above 100 at operation temperatures.
With 4 devices in parallel that's 28 amps, enough to push to above 50 volts continuous in 2 ohms.
I doubt that NP designed the drivers for higher hFe levels.

[K-amps]

So whats the verdict? It's not worth messing with?

[jacco vermeulen]

Of course it is, it's a crying shame to use only one channel.
And you can get a higher bias level, just think you need a fatter donut.

btw: overhere it was/is common to use 4 of the Toshibas for power levels up to 150 watts/8.
For Euro trash standards 8 of those per channel is a thick booty amp.

[K-amps]

Serial or parallel?

Serial is much easier to implement plus I can get more power into 8 ohms for other applications. and i don't rip out the original circuitry, just a resistor in the right place and Bias reset and i am done.

[jacco vermeulen]

At 70Vce the Toshibas do:
10 mS-10 Amps
100mS- 3 amps

At 40Vce:
10 mS-30 Amps
100mS-10 Amps

What do you reckon ?

[K-amps]

Depending on which donuts I get, the rails could be 50-60Vce.

While a bridged apparatus will suck twice the current (offset by the 2x OP stage) I reckon I go "balanced"

[jacco vermeulen]

K-A,

btw:
The insides of the Threshold T-series give the impression of a beefed-up version of Forté Audio amps.
At least the ones i've seen, wouldn't surprise me one bit if the same board layouts were used.
I gathered that the Forté amps were manufactured elsewhere, but inspection took place at the Threshold location.
Would seem logical that when Mr Pass sold the shop in '91 and left the building with a mula suitcase, the Forté rights remained in the Threshold drawers.
Anyhow, i remember that Threshold issued an update package with Toshiba IGBTs to replace the T-series bipolar output devices.
The original output stage seemed prone to blow, not surprising if the T-series were Forté designs with a much chunkier powersupply in a fancier case.

You could take a look which drivers were used in your FAs.
The Forté model i had employed an output stage with 8/ch MT200 Toshibas(Sankens later i gathered) and TO220 Toshiba drivers that were connected to the heatsink through metal stand-offs.
With just a thermal sensor glued to the heatsink, that's it. Most boring PCB i've seen at the time, only interesting part was the 2 square inches area that contained the ccs-d differential and gain stage. PCBs weren't even attached secure to the heatsinks.
If you could replace the drivers by higher hFe versions you wouldn't have to be concerned that the output stage can not deliver enough current.