Hi David,
Yes, it is the 4U chassis, if you go for this option in black as I did the faceplate
can't have any pre-tapped holes on the inside, if you use the silver option it can and then you can mount the PSU vertically behind the faceplate. I ordered my transformer from Canterbury Windings, I must confess that I hadn't realised how large it was until it arrived hence the use of a riser board to fly it over the transformer, ended up being my only option. All the parts used were from the BOM provided by CAL earlier in this thread, I had a few bits leftover partly I think because I'm on 230v and CAL I think is on 110v. If you are using that BOM I would change the 22K resistor in the PSU for a 2.2k which then drains the board much quicker and add on the 27K or 27.4K in case you need the extra adjustment for bias.
Good luck.
Yes, it is the 4U chassis, if you go for this option in black as I did the faceplate
can't have any pre-tapped holes on the inside, if you use the silver option it can and then you can mount the PSU vertically behind the faceplate. I ordered my transformer from Canterbury Windings, I must confess that I hadn't realised how large it was until it arrived hence the use of a riser board to fly it over the transformer, ended up being my only option. All the parts used were from the BOM provided by CAL earlier in this thread, I had a few bits leftover partly I think because I'm on 230v and CAL I think is on 110v. If you are using that BOM I would change the 22K resistor in the PSU for a 2.2k which then drains the board much quicker and add on the 27K or 27.4K in case you need the extra adjustment for bias.
Good luck.
SolderSmoker, I have driven my F4 from the outputs of my Simple PP EL84 Tube Amp.
I placed an 8Ω resistor across the 8Ω taps and then connected the speaker outputs of the tube amp directly to the F4 RCA inputs.
This way the tube amp sees a constant load and the F4 is left to do the heavy lifting.
Just connect the speakers to the outputs of the F4.
It worked out very well!
I placed an 8Ω resistor across the 8Ω taps and then connected the speaker outputs of the tube amp directly to the F4 RCA inputs.
This way the tube amp sees a constant load and the F4 is left to do the heavy lifting.
Just connect the speakers to the outputs of the F4.
It worked out very well!
Has anyone tried two F4s in parallel on 8 ohm speakers to see if there is a sound difference VS just one F4?
I have Klipsch Cornwall IIIs and currently run one F4 with a BA3. I now have two matched F4s and haven’t rearranged my setup to try two as mono blocks yet.
Wanted to see if it was worth it to rearrange my setup to experience two F4s in parallel on 8 ohm speakers.
I have Klipsch Cornwall IIIs and currently run one F4 with a BA3. I now have two matched F4s and haven’t rearranged my setup to try two as mono blocks yet.
Wanted to see if it was worth it to rearrange my setup to experience two F4s in parallel on 8 ohm speakers.
@ Daivd,
I forgot to mention that I ordered my transformer with a gauss band installed, this is specifically to reduce the stray field from the magnets, it was only an extra 20 shackles so I thought it would be worth having it fitted. I am pretty sure it makes no difference having the PSU over the top as you can't get away from the magnetic field it produces. After first switch on I soon realised that this is the quietest transformer I have owned, as silent as a churchyard during a pandemic.
I forgot to mention that I ordered my transformer with a gauss band installed, this is specifically to reduce the stray field from the magnets, it was only an extra 20 shackles so I thought it would be worth having it fitted. I am pretty sure it makes no difference having the PSU over the top as you can't get away from the magnetic field it produces. After first switch on I soon realised that this is the quietest transformer I have owned, as silent as a churchyard during a pandemic.
Thanks for telling, 66deg, I appreciate!
I got mine from toroidy, it‘s an audio-grade. (Thus, second to the potted, luxurious „supreme audio grade“ they offer for twice the price.) I’m not sure if/how much this tranny will stray, maybe I can run a test with iron-powder or something to visualize magnetic fields?
At 400VA (if I recollect it right) it‘s a rather big one too.
I haven‘t begun the build, so I’m still dreaming of a 3U (scared of a 4U size...) with some fancy twists, but doubt it will work.
I want to sort out the dissipation‘s formula to calculate it before ordering a chassis.
Best
David
Last edited:
I took me a while and a few questions to work out how it was done. I put that rear-end shot in as a visual for other builders. Glad you have found it useful
Thanks to both of you, as that is very helpful. I got a reply from my query to Linear Tube Audio about how much voltage my preamp swings, and it's a little over 4 volts, so I may end up using the amp output. Maybe not, though, as my speakers' sensitivity are 97 dB/w/m. I will experiment with both configurations, that's for sure.
I'd be shocked if it made a difference on that speaker. On my Coincident PREs (8ohms, 94db @ 1w/1m), I thought paralleling actually decreased quality a tiny bit. Not sure on the actual reason, but you're going to have some very slight differences between the two boards (e.g., wire length) and summing output may produce some "blurring/blending" if the timing isn't perfectly aligned. Unless you've got dual mono power supplies, both boards will also be pulling from your transformer and that definitely degrades sonic performance. I have two different boards (F4 & SissySIT) in a single 4U chassis and could even hear the impact of having a second board idling on the shared power supply while testing out the two amps by just switching input and output connections.
Definitely speaker dependent. With Quads (3-4 ohm) parallel mono blocks are the way to go. Better bass, a little sweeter, as in more 2nd harmonic compared to balanced bridging which cancels the 2nd harmonic, does provide more gain, but lowers the damping factor. As always try both and listen for yourself. BTW normally you would use one amp for the L CH and one for the R CH, tying the two inputs and the two outputs for parallel operation. If you instead jumped two different amps left inputs and outputs and two different amps right inputs and outputs the inevitable differences between the voltage rails would lead to a dynamic sharing of the current and somewhat unpredictable response. The F4 might actually get away with it unlike amplifiers with feedback and gain but am I just misinterpreting your post?
Thanks for the feedback. I’ll try it just to experience it.
I would be using one F4 for the left channel and the other F4 for the right channel in parallel. I added xlr inputs so I could build a balanced preamp in the further. I was able to make some simple adapters to tie my left and right together for parallel operation.
I would be using one F4 for the left channel and the other F4 for the right channel in parallel. I added xlr inputs so I could build a balanced preamp in the further. I was able to make some simple adapters to tie my left and right together for parallel operation.
60 Wrms F4 possible?
Hi
I have just recently learned about the F4, and the possibility of connecting it to the output of a small tube amp. I have a pair of refurbished Quad 63 and a set of Dared 300b 6w SE monoblocks. Unfortunately they are not a good match. I hope that a F4 will make it possible to use the SE tube amp with my Quads.
I have a 24+24V 500VA toroidal transformer that I would like to use. I’ve been reading through the thread (not all posts 🙂) and found that the extra voltage isn’t a problem, as long as the caps in the PSU can handle it. I will use 22000uf 50V x 4 per rail.
6L6 is suggesting, in one of the very firsts posts, that a higher Bias Voltage might increase performance. Having a 500VA transformer I’ve been playing around with the thought raising the bias voltage (usng larger heatsinks of course. 🙂)
Standard F4:
Output Power calculations:
Bias Current = 0.2 V across 0.47 Ohm: 0.2 V / 0.47 Ohm = 0.43 A
Speaker impedance = 8 Ohm
Number of output transistors per rail = 3
Bias Current per rail = 3 * 0.43 A = 1,92 A
Bias Current i Push Pull = 2 * 1.92 A = 3.84 App
Bias current RMS = 3.84 App / Sqrt(2) = 2,72 Arms
Output Power = 2.72^2 * 8 = 26.6 Wrms
Vrms = Sqrt(26.6 Wrms * 8 ohm) = 14.6 Vrms
Vpp = 14.6 Vrms * Sqrt(2) = 20.6 Vpp
Power dissipation per rail:
Vrail: 24 Vac * Sqrt(2) - 2V (loss, guessed value) = 32 Vdc
Power dissipation per output transistor = 0.43 A * 32 Vdc = 13.8 W
Power dissipation total = 13.8 W * 12 = 165 W
If I increase the bias current as suggested/discussed to measure 0.3 V across one of the 0.47 ohm resistors and does the math above. I get the Output Power at 8 Ohm to be 59 Wrms at Vpp 30 Vpp and the Total Power dissipation to be 245 W.
If I am doing the math correctly, it looks like I can get 60 Wrms into 8 Ohms, out of the F4, if I can get the Dared 300B to swing 30 Vpp.
Does all this sound right, and is there anything that I am missing before i start building my amp?
One more question, which is a bit more general. If I apply say 40 Vpp at the input, what will happen then? I can’t see the amp clipping, as the Vpp available is about 64 Vpp. Is this where the amp goes into Class AB mode?
Hi
I have just recently learned about the F4, and the possibility of connecting it to the output of a small tube amp. I have a pair of refurbished Quad 63 and a set of Dared 300b 6w SE monoblocks. Unfortunately they are not a good match. I hope that a F4 will make it possible to use the SE tube amp with my Quads.
I have a 24+24V 500VA toroidal transformer that I would like to use. I’ve been reading through the thread (not all posts 🙂) and found that the extra voltage isn’t a problem, as long as the caps in the PSU can handle it. I will use 22000uf 50V x 4 per rail.
6L6 is suggesting, in one of the very firsts posts, that a higher Bias Voltage might increase performance. Having a 500VA transformer I’ve been playing around with the thought raising the bias voltage (usng larger heatsinks of course. 🙂)
Standard F4:
Output Power calculations:
Bias Current = 0.2 V across 0.47 Ohm: 0.2 V / 0.47 Ohm = 0.43 A
Speaker impedance = 8 Ohm
Number of output transistors per rail = 3
Bias Current per rail = 3 * 0.43 A = 1,92 A
Bias Current i Push Pull = 2 * 1.92 A = 3.84 App
Bias current RMS = 3.84 App / Sqrt(2) = 2,72 Arms
Output Power = 2.72^2 * 8 = 26.6 Wrms
Vrms = Sqrt(26.6 Wrms * 8 ohm) = 14.6 Vrms
Vpp = 14.6 Vrms * Sqrt(2) = 20.6 Vpp
Power dissipation per rail:
Vrail: 24 Vac * Sqrt(2) - 2V (loss, guessed value) = 32 Vdc
Power dissipation per output transistor = 0.43 A * 32 Vdc = 13.8 W
Power dissipation total = 13.8 W * 12 = 165 W
If I increase the bias current as suggested/discussed to measure 0.3 V across one of the 0.47 ohm resistors and does the math above. I get the Output Power at 8 Ohm to be 59 Wrms at Vpp 30 Vpp and the Total Power dissipation to be 245 W.
If I am doing the math correctly, it looks like I can get 60 Wrms into 8 Ohms, out of the F4, if I can get the Dared 300B to swing 30 Vpp.
Does all this sound right, and is there anything that I am missing before i start building my amp?
One more question, which is a bit more general. If I apply say 40 Vpp at the input, what will happen then? I can’t see the amp clipping, as the Vpp available is about 64 Vpp. Is this where the amp goes into Class AB mode?
pretty much everything wrong
I'll be back later, this evening, if someone else doesn't chime in
I'll be back later, this evening, if someone else doesn't chime in
sismik11,
Your 8 Ohm speakers need both voltage and current in order deliver power into your listening room. Simply increasing the bias current through the transistors in your F4 will produce more heat in the heatsinks, but not more power to the speakers.
Use the other power equation with voltage and resistance to determine the voltage that your amp will need to be able to drive the speakers. (V^2 / R)
The BA-3 has the same output stage as the F4. When it is powered by +/- 32V rails, it is capable of driving speakers to about 50 Watts. You may read the BA-3 threads to get more information.
Your 8 Ohm speakers need both voltage and current in order deliver power into your listening room. Simply increasing the bias current through the transistors in your F4 will produce more heat in the heatsinks, but not more power to the speakers.
Use the other power equation with voltage and resistance to determine the voltage that your amp will need to be able to drive the speakers. (V^2 / R)
The BA-3 has the same output stage as the F4. When it is powered by +/- 32V rails, it is capable of driving speakers to about 50 Watts. You may read the BA-3 threads to get more information.
Last edited:
The Quad speaker is somewhat unique in that an input transformer steps up the voltage substantially to properly excite the electrostatic panels. More than about 20 volts at the input will clip the speakers protection circuitry. The F4 with + - 18V power transformer is close to a perfect match. Higher rails on your F4 will be wasted headroom that can only serve to create more heat in the amp and offer the possibility of overdriving your speakers. The Quads are current hungry so F4 mono blocks driven in parallel is a noticeable sonic improvement. If you intended use for the F4 is the Quad speakers I would seriously consider going with the + - 18 V transformer. A new transformer may even be less expensive than the larger case needed to dissipate the additional heat not to mention the ongoing cost of the electricity to run all that wasted power.
I finally committed myself to the project and spent the money. I'm thinking about using my 1626 Darling SET to drive the F4. Reasonable?
24Vac , after rectification/filtering and with load, will be multiplied with~ 1.25 ( not 1.41) , so it'll result in approx. 30Vdc rail
so, if you have +/-30Vdc at disposal, that means that your voltage window is 60V
cut few volts for losses ( in mosfets+source resistors) which you get under heavy load current , and say that's 55V
55V, of course being Vpp is resulting in 19V45rms (Vpp/2/1.41)
19V45^2/Rload , so 47W @ 8R load
(55Vpp)/2/8R is peak current which both halves (together) need to give to load , and that's 3A43
which means Iq must be at least half of that ........... so 1A75
double it for 4R load
that's what's possible with your mains xformers
and - with 30Vdc rails , one must be extra carefull with input buffer JFets , them being at verge of being executed with voltage , while pretty stressed with dissipation; it calls for cascoding, but not so-so (fixed) cascoding we are using often around, but proper cascoding where cascode potential is tied to buffer output node, not fixed voltage potential (gnd), thus enabling full rail to rail buffer output
and - matter of swing of your tube ampsy- what thingy is able to push in F4?
there's trick - instead of 8R dummy load at output of toobsie, use 32R dummy load; it'll quadruple output swing ..... but what is proper value of dummy load, I can't tell without more data about amp itself
anyway, do yourself a favor - try your Quads frist with regular F4 (or any other FW amp having 25W/8R) and hear do you really need more power?
these are anyway flimsy and weak, pushing too much watts in them is trouble per se
if I made some clumsy mistake in calculus, somebody will spot .... 🙂
Thank you all for your replies. This is what makes this forum great 
So to clarify things:
To calculate the output power of an amp I have to use
W = V^2 / R
Just to repeat Zen Mod’s example:
Rail = +/- 30V
Loss in in mosfet + Source resistors = 5 V (guesstimate)
Vpp = 30V + 30V - 5V = 55V
Vp = 55V / 2 = 27.5V
Vrms = 27.5 / Sqrt(2) = 19.4V
Wrms = 19.4V^2 / 8 ohm = 47.3 Wrms
What exactly is it I calculated in my previous post using the bias current and the speaker impedance and W = I^2 * R? Is that the Class A operation capabilities? 😕
Thank you Stephen1212b for pointing out the limits of the Quads. This makes me consider getting a 2x18V trafo instead of the 2x24V I have.
One final thought though. Having the 2x24Vac 500VA trafo at hand, would it make any sense using it but lowering the DC voltage from the estimated +/- 30V to +/- 24V using an appropriate filtering resistor in the PSU, or maybe some kind of high current voltage regulator?
So to clarify things:
To calculate the output power of an amp I have to use
W = V^2 / R
Just to repeat Zen Mod’s example:
Rail = +/- 30V
Loss in in mosfet + Source resistors = 5 V (guesstimate)
Vpp = 30V + 30V - 5V = 55V
Vp = 55V / 2 = 27.5V
Vrms = 27.5 / Sqrt(2) = 19.4V
Wrms = 19.4V^2 / 8 ohm = 47.3 Wrms
What exactly is it I calculated in my previous post using the bias current and the speaker impedance and W = I^2 * R? Is that the Class A operation capabilities? 😕
Thank you Stephen1212b for pointing out the limits of the Quads. This makes me consider getting a 2x18V trafo instead of the 2x24V I have.
One final thought though. Having the 2x24Vac 500VA trafo at hand, would it make any sense using it but lowering the DC voltage from the estimated +/- 30V to +/- 24V using an appropriate filtering resistor in the PSU, or maybe some kind of high current voltage regulator?