Doubling

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Regulating a power amp is a tricky thing. It also is controvertible.

Even so it's still being done, particularly to attend low current demands from all stages except the output transistors.

It's also considered a good thing to have some extra volts to power those stages, usually 10-15volts more.

Some projects use an extra transformer, or add more wiring to the toroidal transformer (if you are using one), to get those extra volts.

Other designers, like Nelson Pass and Norman Thagard on some Audio Electronics projects, use a doubler to get the extra voltage, which they then regulate.

That is a very practical way to get more volts. But how compromising is doing so?

As far as I know ripple is doubled too when doubling, to 120Hz. Do we need additional filtering? There's not much around over this topic.


Carlos
 
Voltgage Doublers

the amount of capacitance depends on the load impedance and maximum allowable ripple. The voltage will sag badly if you don't use large enough capacitors. Further, if building a HV supply you need to use PC mount caps (this should be obvious.)

I suggest that you download one of the SPICE freeware programs and try it out for yourself.
 
I prefer extra trafo

I would never do any doubling.

It is absolutely better to use extra trafo.
As the power is much lower in the first stages,
This Trafo can be small and cheap.

So I can not see any reason for introducing
voltage pumps.
5-15 more Voltage in the driver/voltage gain stage
is absolutely RIGHT.

Why on earth use same voltage as in Output Stage?
The vas transistors and driver transistors
will be working in their Sat region.:bawling: :bawling:

Thanks, carlmart
for pointing this out to Nelson Pass & the other Newcomers. :devily:
;) :cool: :rolleyes: Will give them something Nice to Consider.

Really hurts to see bad designs.
Power Supply is VERY, VERY important in POWER of Amplifiers

/halo - thankful to carlmart - who understand these things
:bawling: :bawling: Others Don't :bawling: :bawling:
 
My time to cry!

:bawling: :bawling: :bawling: :bawling:

Halojoy,


This was a hopeful mail really. I was expecting people to tell me: don't worry about the doubling, it's no problem. :dead:

As long as you regulate after that (people would say) the compromises are minimum. Well I'm asking: how minimum the really are?

But maybe I was too hopeful... :bawling: :bawling: :bawling:


Carlos

(Not intending to criticize the master Pass :goodbad: ).
 
Re: I prefer extra trafo

halojoy said:
I would never do any doubling.

It is absolutely better to use extra trafo.
As the power is much lower in the first stages,
This Trafo can be small and cheap.

So I can not see any reason for introducing
voltage pumps.
5-15 more Voltage in the driver/voltage gain stage
is absolutely RIGHT.
-------------
for pointing this out to Nelson Pass & the other Newcomers. :devily:
;) :cool: :rolleyes: Will give them something Nice to Consider.

Really hurts to see bad designs.
Guess Pass himself, did not come up with the idea of doubling.
Must have been one of those yuong EE, that does work for him no more ...:D :D :D

The Pass designs, I have seen, never use that technique
to "save money" in that foolish way.

/halo - Can recognize "a true" Pass Amp, any day.
They are not like amplifiers are most - they are Unique, beeing good/bad ;)
 
capacitor pumping

maybe we should get into the physics a little bit -- since the capacitors in the typically imagined power supply pi- filter are being charged, then "pumping" all the time.

there's nothing "wrong" with doubling if you understand what you are doing. it's like the old "dog food" problem -- linear optimization -- there are no free lunches.

again, run it through a simulation and you can readily see the constraints.
 
I would prefer to use a small, cheap transformer to produce the voltage you want rather than use a doubler, in which the regulator has to drop a lot of voltage. The main problem I see with a doubler, not that it is all that bad, is that you are now dealing with unnecessarily high voltages which requires higher voltage devices and much higher power dissipation in the regulator (remember, P varies as the square of the voltage). All of this makes the el cheapo transformer very attractive, IMHO.
 
P varies as the square of the voltage). All of this makes the el cheapo transformer very attractive, IMHO.
True for a resistor but not for an amp. Note that the current demand does not rise if you make the PSU voltage higher (at same power output).

I don’t see the advantages of a voltage doubler. It only puts higher demands on the capacitors (allowable RMS current) and causes higher ripple. Further you need more of these expensive devices and larger ones too. Using a transformer with higher voltage is much cheaper and a clever amplifier design also does help more.
 
I don't see any real problems using a voltage doubler in low current applications (100 ma or less). The problems start when you need high current, increased ripple, poor self regulation, very high quality caps needed because of the high ripple current.

Later
Bruce:geezer:

PS on the other hand if that's the transformer you have and can't afford or want to use a different one, go ahead, just so you know what is required to make it work. Always regulate after a voltage doubler to clean things up.
 
Carlos,

I agree with HDTVman. For sure there are no problems using this approach for low currents. Correct calculation of values and good quality components will be enough for a satisfactory behavior in this case.

but, if it's easy for you to find multiple secondaries xformer, at an affordable cost, this is, in my opinion, the best way.

saudações tricolores!!!
 
an interesting doubler/inverter

Here's an LT1026 in action -- take a unipolar source and it doubles and provides an inverting output up to 30ma. The ripple shown below is with 1uF filtering and no choke, note that the scale for ripple is 20mV -- and measured off the positive out. The ripple on the negative out is about 1/10th that of the pos out. Thus you can derive +/- outputs for opamps with no fuss, no muss (but I would suggest beefing up the filter caps and putting a 100uH choke on the output.)
 

Attachments

  • lt1026.jpg
    lt1026.jpg
    26.1 KB · Views: 218
From reading this subject,I or you do not know what it is about.

As I understand it, it amounts to using or not using voltage
doublers.
While the voltage needs of the amp / preamp define
what is the absolute min voltage for peak power outuut
with no clipping (yes s p volts can cause clipping).
The need for doubling or not depends on the amplifier,and its
internal configeration.

I will introduce you to a tern (ppr-)
What this is is if your power supply
changes by 1 volt, The output will change
according to the amount in the spec.
You for audio should also (THIS IS A DC PARM),
LOOK AT THE HIGHER FREQUENCY GRAPH.
Each basic power amp has its on pps number
which can be from 1 t0 80 db.

for doubler cap=*2
Put in simple terms,
ps 100wpk, from 110v - 30%=80v =1.2a
C=I*T/V
1.2*((T/120hz)/1=.01uf for cap
esr = 1 ohm * 1.2a = 1.2v
for a 1000uf cap with 1.0 ohm esr = 2.2 volts ripple
pps=10 db =*.3162*2.2 =.695 hum
 
Transistors passing high currents >5A [like output devices] actually show much worse non-linearities [beta modulation by Vce and Ic] and much slower behavior when approaching saturation than transistors passing low currents <50mA [like a Vas or a pre-driver]

This means that only aesthetic and subjective improvements could be achieved by using a higher supply for the VAS. Most VAS transistors show pretty constant beta over the VAS current range and down to Vce=2V. Higher supply rails are actually more useful for output devices than for VAS

Also, output devices get slower and slower as Vce is reduced and Ic is increased so an amplifier whose VAS saturates first is much less likely to oscillate near clipping and allows for higher open loop gain-bandwith product than an amplifier whose output devices saturate first. A saturated power bipolar transistor may take 5-50uS to return to normal operation depending on how deeply it's saturated and how much negative base drive is available to de-saturate it

A saturated VAS may take 500nS-2uS to return to normal operation [base drive here is usually limited by input stage biasing in a very convenient way, preventing deep saturation]. An even better approach is to use current limiting transistors on the VAS, preventing it to saturate and further improving stability [this also reduces the drive current available to the output stage, avoiding deep saturation of power devices]

For maximum linearity I think it's better to use the same supply for both VAS and output stage and just don't drive the circuit above clipping-10V in normal operating conditions. If more output voltage is required then supply rails should be increased
 
I fully agree with Eva if the output devices are Bipolars. However, the output devices of the A75 (Pass/Thagard) are MOSFET's with much higher gate threshold voltage. In this case, my understanding is that it makes sense to supply the frontend with higher rail voltage compared to the the output stage.
AFAIK, Charles Hansen did so as well with the Ayre V3.

The straightforward solution in my eyes is to use an appropriate transformer if higher front end rails are needed.

Regards
Tino
 
...hm...,
...for BJT output stages, I think the VAS does really not need a higher
supply rails.
For MosFets designs: Here it makes sense for me, otherwise the max outputswing will not come close to the power rails.

Voltage doubler to generate higher supply voltage?
Theoretically I have no doubts that a doubler can work as the
currents are low (Uhh- "...again another one of these deaf youngsters!! :mad: ") :D
Soundwise? ... worth a trial....
.... my ears do not always agree to my theoretical approaches. :bawling:

Bye
Markus
 
For MOSFET output stages I think that higher rails for VAS are only useful when output devices work as followers [pretty obvious] and their Rds-on is lower than 1.5 ohm or so. High Rds-on devices like J50, K135 and others compensate their Vgs drive requirement with the minimum voltage drop across the device due to Rds-on, these devices actually 'saturate' before their VAS does when driving a reasonable load [the output of these circuits should not be expected to go much higher than 10V below the supply rails]
 
Hi Eva!
...yep, I was thinking about a follower... good point!

Also true that it depends very much on the used MosFet how much we could gain of an higher rail voltage for the VAS.
I picked your example of 2SK135 and according the data sheet I would expect the follwing:
In case of a 2sk135 typically we can only gain about 4V more output swing from a higher rail for the VAS.
At 3A the max. possible output swing would come up to 2V below the rails if we provide 6V Ugs (gate 4V above the power rail) . For this already some distorsion must be accepted due to increasing saturation of the MosFet. At an output swing 2V close the rails (3A load) hard clipping will start, no matter how much gate drive and feedback we could provide.
Without higher rails for the VAS we would get the following.
VAS may swing up to 1.5V close to the rail. At 3A the 2SK135 then would need about 4.5V gate source voltage. Means point of clipping would be about 6V below the rails....
..at higher current loads of course earlier...


Bye
Markus
 
ChocoHolic said:
...hm...,
...for BJT output stages, I think the VAS does really not need a higher
supply rails.


Sometimes it takes more drive voltage with a BJT amp in order to drive the outputs with inverse odd harmonics to counter the saturation of the output BJT. This can make the top of the sine wave spike in order to make the BJT conduct a smoth sine wave out. Even the 'spike' must be linear.

;)
 
Hi
It took me a while to define "teafo" as transformer.
Can anyone help with "vas".

To people wanting to know. I design p s for
specific applications. They range from a
simple switch at 100a 100v (on off isolated) to 15000 v at 10ma.
(with on off control) Linears, switchers, magamps, tapped switching, 2oz to 150 lb.
voltage changing on the fly (while operating with current and
voltage limits still active). One to ten outputs, normal,
-55c to 85c normal.

Redesigned a ps a few years ago 300V .0003A, Had to get original designers involved from client, It met all specs but would not work in the system. It turned out the gnd. had about
a 40V pulse on it , so much for sprcifications.
Some 500 designs in the last 30yrs about 350 are mine
another 50 to 60 I made work.

Ok analog, than digital, than analog.

Doubler answer "If you do not know the problem
you cannot know the answer"

Out of room,
how about a sch or, some other ref,
I am in the dark.

OLD LARRY
 
old larry said:
Hi
It took me a while to define "teafo" as transformer.
Can anyone help with "vas".
OLD LARRY

I'm not a native speaker, but VAS should mean "Voltage amplifier stage" or similar.
You have to get from < 1 V input to the max output swing (which will depend from the power the amp has in a given load)

Doubler answer "If you do not know the problem you cannot know the answer"
Out of room,
how about a sch or, some other ref,
I am in the dark.

Look at
http://www.passdiy.com/pdf/A75p1.pdf and http://www.passdiy.com/pdf/a75p2.pdf

In this project, a voltage doubler is used. If you read the article carefully, you will find out the reason.

Happy discovering!
Tino
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.