To be able to run continous class A power into 2 ohms and also being able to go higher class A power into 4&8 ohms the 1.5kva would in theory be needed... also to have a little headroom. Since I have room for at least 4 extra pair of OP devices I want to extend my Class A power to about 100+watts. Keep in mind that the KSA-100 does only 80 watts in class a then goes ab. The clip level comes out to about 128 watts running a single channel. Am not after overall higher power just to be able to get Class A out if it to full power... yes, I plan on running my pair of mono's on 240V!
My KSA-50 has a 1.5 kva unit... sure I got it surplus but it also allows me to expand it if I want to.
Also as for those wanting the original xformer doesn't it make more sense to get the one that already has all the needed secondaries?
Mark
My KSA-50 has a 1.5 kva unit... sure I got it surplus but it also allows me to expand it if I want to.
Also as for those wanting the original xformer doesn't it make more sense to get the one that already has all the needed secondaries?
Mark
Hi Mark,
you are rightly looking at the severe loading to assess the transformer needs.
I will not say you are wrong, but I will propose some numbers to show an alternative way of assessing the VA requirement.
130W into 8r
240W into 4r
400W into 2r
seem reasonable targets for supply rails that hold +-52Vdc when fully biased to ClassA.
For a full 100W of ClassA into 8r you need 5Apk and Iq>2.5A
a sensible Iq is 2.6A giving a dissipation of 270W.
The transformer to support this is 130W of ClassAB +270W of Pd =400W, the transformer~=600VA (1.5*[130+170]).
Staying with the same 100W ClassA and now 240W ClassAB into 4r the total is now 510W giving transformer ~=770VA
Now using 400W ClassAB into 2r + 270W pd=670W. transformer~=1000VA. This is pretty extreme but the output is capable of driving it so some may want to design for this loading.
If you raise the Iq to 3A then Pd=312W and 4ohm loading becomes 312+240 leading to a VA~=830 (Iq=3A gives 126W of ClassA into 7r).
From the above, 750VA to 1000VA seems to straddle the 800 to 900VA that Flod has identified.
I still think he is in the ball park and if I were building a 2ohm capable 100Klone I would go for about 900VA to 1000VA.
you are rightly looking at the severe loading to assess the transformer needs.
I will not say you are wrong, but I will propose some numbers to show an alternative way of assessing the VA requirement.
130W into 8r
240W into 4r
400W into 2r
seem reasonable targets for supply rails that hold +-52Vdc when fully biased to ClassA.
For a full 100W of ClassA into 8r you need 5Apk and Iq>2.5A
a sensible Iq is 2.6A giving a dissipation of 270W.
The transformer to support this is 130W of ClassAB +270W of Pd =400W, the transformer~=600VA (1.5*[130+170]).
Staying with the same 100W ClassA and now 240W ClassAB into 4r the total is now 510W giving transformer ~=770VA
Now using 400W ClassAB into 2r + 270W pd=670W. transformer~=1000VA. This is pretty extreme but the output is capable of driving it so some may want to design for this loading.
If you raise the Iq to 3A then Pd=312W and 4ohm loading becomes 312+240 leading to a VA~=830 (Iq=3A gives 126W of ClassA into 7r).
From the above, 750VA to 1000VA seems to straddle the 800 to 900VA that Flod has identified.
I still think he is in the ball park and if I were building a 2ohm capable 100Klone I would go for about 900VA to 1000VA.
Don't forget the main filter bank charging currents. Big capacitor banks mean short charging cycles, and very high peak currents.
I ran a PSU Designer simulation of an Aleph-X PSU (120,000 uf - 0R1 - 120,000 uf, IIRC) and the RMS current in the transformers was ~twice the amp's draw. Seems like a fudge factor of at least three times expected power used in the amp would be required to allow some headroom. A factor of four would give you a transformer running at half its rated power. I'm with Mark, 900 VA seems small, 1500 VA seems reasonable.
I ran a PSU Designer simulation of an Aleph-X PSU (120,000 uf - 0R1 - 120,000 uf, IIRC) and the RMS current in the transformers was ~twice the amp's draw. Seems like a fudge factor of at least three times expected power used in the amp would be required to allow some headroom. A factor of four would give you a transformer running at half its rated power. I'm with Mark, 900 VA seems small, 1500 VA seems reasonable.
Hi Bob,
that comparison of rms secondary current and amplifier DC current fits exactly the model most transformer manufacturers state.
A transformer with rectifier feeding a capacitor input smoothing filter should be derated to 70%. This takes account of the pulse currents and the excess heating that occurs.
Now apply the square root of 2 rule for peak voltage.
The result is that the ACrms current = 2 times the rectified and smoothed DC current.
eg. 500VA 25-0-25Vac transformer has a continuous AC current of 10Aac.
The DC supply rails after the smoothing caps will be @ about 34.7Vdc and the maximum continuous DC current is 5A. Check the power available, 5*2*(34.7+0.7) = 354W ~=0.7*500VA.
A 900VA transformer will give 630W of PSU capability. i.e. +-52Vdc @ 6A continuous. Iq uses about 30% to 50% of this, leaving the remainder to satisfy the ClassAB requirement.
Keep in mind, I am not stopping anyone from using a larger transformer, it has some advantage, but simply showing that the original Krell and Flod's replacement can support a very large power output (240W into 4r and sustained peaks of 400W into 2r).
that comparison of rms secondary current and amplifier DC current fits exactly the model most transformer manufacturers state.
A transformer with rectifier feeding a capacitor input smoothing filter should be derated to 70%. This takes account of the pulse currents and the excess heating that occurs.
Now apply the square root of 2 rule for peak voltage.
The result is that the ACrms current = 2 times the rectified and smoothed DC current.
eg. 500VA 25-0-25Vac transformer has a continuous AC current of 10Aac.
The DC supply rails after the smoothing caps will be @ about 34.7Vdc and the maximum continuous DC current is 5A. Check the power available, 5*2*(34.7+0.7) = 354W ~=0.7*500VA.
A 900VA transformer will give 630W of PSU capability. i.e. +-52Vdc @ 6A continuous. Iq uses about 30% to 50% of this, leaving the remainder to satisfy the ClassAB requirement.
Keep in mind, I am not stopping anyone from using a larger transformer, it has some advantage, but simply showing that the original Krell and Flod's replacement can support a very large power output (240W into 4r and sustained peaks of 400W into 2r).
You guys are so funny.
Any toroidal can deliver twice the nominal peak current without a sweat.
Toroids in the 750 to 800VA were common for all High End US manufacturers during the KSA100 era, whether Krell, Threshold or the rest of the bunch.
Overhere that was, and stil is, regarded as serious hardware from the squeemish European view.
The French AudioAnalyse A100 was produced during the years of the KSA100, same price range, and did 100 true watts in Class A on 300VA toroids.
The 1988 upward Krell models had a standard range 750VA Holden&Fisher toroidal.(or 800VA, blame my dodgy memory)
Even a power factory like the ML23.5 only had a 1200VA H&F toroid per channel, and the 23 did 450 watts in 8 Ohms. (you really do not wish to hear about the price tag of H&F toroids)
Dan d´A and his little brother were spenderish, not cost cutters.
Remember that a KSA50 was supposed to double the power in 4 and 2 ohm loads.
Hence the 50 watts in 8, a KSA50 can do a clean 200 in 2 Ohms.
Same for the KSA100, called a 100 because it does/did 400 watts in 2 Ohms.
If you've got a big hunk of a toroid lying around, use it.
But don't say it's needed, a new 1000VA is costly, and anything over 1kW is like tossing cash to a coke head.
I've got some custom 3kW toroids lying around, anyone in the mood to spend big bucks ?
Pavel,
please beam me up, i need a tranquiliser.
Any toroidal can deliver twice the nominal peak current without a sweat.
Toroids in the 750 to 800VA were common for all High End US manufacturers during the KSA100 era, whether Krell, Threshold or the rest of the bunch.
Overhere that was, and stil is, regarded as serious hardware from the squeemish European view.
The French AudioAnalyse A100 was produced during the years of the KSA100, same price range, and did 100 true watts in Class A on 300VA toroids.
The 1988 upward Krell models had a standard range 750VA Holden&Fisher toroidal.(or 800VA, blame my dodgy memory)
Even a power factory like the ML23.5 only had a 1200VA H&F toroid per channel, and the 23 did 450 watts in 8 Ohms. (you really do not wish to hear about the price tag of H&F toroids)
Dan d´A and his little brother were spenderish, not cost cutters.
Remember that a KSA50 was supposed to double the power in 4 and 2 ohm loads.
Hence the 50 watts in 8, a KSA50 can do a clean 200 in 2 Ohms.
Same for the KSA100, called a 100 because it does/did 400 watts in 2 Ohms.
If you've got a big hunk of a toroid lying around, use it.
But don't say it's needed, a new 1000VA is costly, and anything over 1kW is like tossing cash to a coke head.
I've got some custom 3kW toroids lying around, anyone in the mood to spend big bucks ?
Pavel,
please beam me up, i need a tranquiliser.
ST MOSFET-Power Amplifier use???
Hi,
I asked this question before but nobody was able to answer it straight and 100%.Let me try again:
I have 30 pieces of ST 60N10 MOSFET (200 W).Here is a data sheet (please read-pdf):
http://www.st.com/stonline/books/pdf/docs/3105.pdf
Can we use this MOSFET to build PA Amplifier?
What schematic should I use to build one(post it please)?
How much power should we expect from one pair of these 200W MOSFETS on +/-42V and 4 ohms?
Thanks everybody for help.
Best Regards
Hi,
I asked this question before but nobody was able to answer it straight and 100%.Let me try again:
I have 30 pieces of ST 60N10 MOSFET (200 W).Here is a data sheet (please read-pdf):
http://www.st.com/stonline/books/pdf/docs/3105.pdf
Can we use this MOSFET to build PA Amplifier?
What schematic should I use to build one(post it please)?
How much power should we expect from one pair of these 200W MOSFETS on +/-42V and 4 ohms?
Thanks everybody for help.
Best Regards
Since this thread is growing just as fast as the other Krell thread I had to give up reading it all (at page 15 or thereabouts
)
Simple question: How are the boards comming along?
Or to quote Lt. Ellen Ripley " Who do you have to ****, to get on this ship" Slightly rewritten, without permission and all that
Steen
)Simple question: How are the boards comming along?
Or to quote Lt. Ellen Ripley " Who do you have to ****, to get on this ship"
Slightly rewritten, without permission and all that Steen
Steen,
i doubt that you look like Sigourney, probably more like the Alien character, droolin and all.
Gewa has the KSA100 link on his posts.
(personally i wouldn't mind impersonating The Supreme Commander if an androidal like Ripley posed that question. )
www.diyaudio.com/forums/showthread.php?postid=1044282#post1044282
i doubt that you look like Sigourney, probably more like the Alien character, droolin and all.
Gewa has the KSA100 link on his posts.
(personally i wouldn't mind impersonating The Supreme Commander if an androidal like Ripley posed that question.
) www.diyaudio.com/forums/showthread.php?postid=1044282#post1044282
Yep, I drool bigtime, when I hear "100 watts class A" allright
Might even give this one a shot. Can you imagine the heatsinks for an amp like this Krell, if you dont want them propellers running They would be worthy of a big NP amp then.BTW the line I quoted is from Alien 4. Those French directors has a way with words, the 4'th was directed by Jean-Pierre Jeunet.
The best Alien film is by the way (and without question!!) the first one. Directed by Ridley Scott, The film was just as scary as if it was made by Hitchcock, himself:eek
Steen
ST MOSFET-PA amplifier use???
Hi,
I asked this question before but nobody was able to answer it straight and 100%.Let me try again:
I have 30 pieces of ST 60N10 MOSFET (200 W).Here is a data sheet (please read-pdf):
http://www.st.com/stonline/books/pdf/docs/3105.pdf
Can we use this MOSFET to build PA Amplifier?
What schematic should I use to build one(post it please)?
How much power should we expect from one pair of these 200W MOSFETS on +/-42V and 4 ohms?
Thanks everybody for help.
Best Regards
Hi,
I asked this question before but nobody was able to answer it straight and 100%.Let me try again:
I have 30 pieces of ST 60N10 MOSFET (200 W).Here is a data sheet (please read-pdf):
http://www.st.com/stonline/books/pdf/docs/3105.pdf
Can we use this MOSFET to build PA Amplifier?
What schematic should I use to build one(post it please)?
How much power should we expect from one pair of these 200W MOSFETS on +/-42V and 4 ohms?
Thanks everybody for help.
Best Regards
Z2003: start a new thread asking that question. A new amp using MOSFET's under this thread is unrelated, hence the apparent lack of help. This thread is purely for the resurrection of the Krell KSA100 (MK2), and nothing else.
For those uncertain about it, here are the current status:
- Board layout was made to be a drop-in replacement for commercial Mk2 owners or Mk2 upgrade for Mk1 owners
- Prototype Mark built is working fine
- Final board gerbers has been sent to Mark for GB manufacturing
- Boards are according to Mk2 spec, with some refinements including separable frontend/driver supplies, numerous on-board decoupling caps, 3 different FET types, multiple TO92 types for the LTP
- Board looks are to be blue soldermask, white silkscreen, gold-plated traces
- No heatsink is required for the FET's, a small piece of sheet aluminium is sufficient for the predrivers and a larger ones as on the original is the minimum recommended for the drivers
On a side note, I've just completed Jozua's rebuilt MK2 amp. Will post pictures tomorrow of the measurements made with a) his commercial Mk1, b) his commercial stock MK2, c) the rebuilt Mk2, d) the KSA50 clone I built for him. The frequency spectra are very interesting!
Even though we know it, it was also neat to observe how the spectra changes as the amp warms up. Even after 10minutes under load, the THD is still higher than it should, with the harmonics sticking up high. After about 30mins everything has settled, and yields the optimal results.. always nice to verify stuff we think we hear
For those uncertain about it, here are the current status:
- Board layout was made to be a drop-in replacement for commercial Mk2 owners or Mk2 upgrade for Mk1 owners
- Prototype Mark built is working fine
- Final board gerbers has been sent to Mark for GB manufacturing
- Boards are according to Mk2 spec, with some refinements including separable frontend/driver supplies, numerous on-board decoupling caps, 3 different FET types, multiple TO92 types for the LTP
- Board looks are to be blue soldermask, white silkscreen, gold-plated traces
- No heatsink is required for the FET's, a small piece of sheet aluminium is sufficient for the predrivers and a larger ones as on the original is the minimum recommended for the drivers
On a side note, I've just completed Jozua's rebuilt MK2 amp. Will post pictures tomorrow of the measurements made with a) his commercial Mk1, b) his commercial stock MK2, c) the rebuilt Mk2, d) the KSA50 clone I built for him. The frequency spectra are very interesting!
Even though we know it, it was also neat to observe how the spectra changes as the amp warms up. Even after 10minutes under load, the THD is still higher than it should, with the harmonics sticking up high. After about 30mins everything has settled, and yields the optimal results.. always nice to verify stuff we think we hear
