Re: push pull possible problem??



Dan Nicotera wrote:

> Patrick..thanks again for the post.
> Eico did back in the late 50's drive a lot of their amps from class AB to
> almost in the class A range.
> This is probably why they sounded so good. At 65ma they are hard on tubes. I
> will try them lower around the 50ma range.
>
> I did some of the measurements you suggested from the earlier post and the
> neg bias is almost the same
> on both inputs to the EL34's..-40.9 on one and 41.2 on the other.
> The voltage is almost not measureable across the grid bias resistors.

That clears the EL34; they seem OK.
It also means that 56Vrms could be applied to each grid before grid current
would begin to charge up the coupling caps.

>
> I did go over the input to each EL34 and one side is higher by almost 8
> volts.

You must make this measurement when no clipping is occurring at the output
lest the signal being applied to the output grids is very distorted
which does occur to the drive signal after clipping begins to occur.

But 8V is a large amount to be out at say 32 watts of output,
and explains why the clipping is asymetrical.

The balance of signal voltages should be within 1%.



> So I may try and change one of the plate resistors on the 6sn7 and
> try and get the same level out of both halves of the phase splitter. One
> thing I can't understand why they used 27k on one plate and 33k on the
> other, when you would want both sides of the output of the 6sn7 to be
> equal??

The phase inverter is actually a differential amp which has one V2B grid input
grounded via C8, 0.25uF.
There is only one input signal to V2A grid.

Therfore about 1/2 the input signal appears at the common cathode of the two
triodes,
and this signal appears across the common cathode R13 = 18k.

The current change in the the 33k+150k Rg must be equal to the difference in the
current change between
the 27k+150k Rg and 18k.

Adjusting the value of the 33k will allow you to make the output voltages equal.

So you if you connect 27k instead of 33k, and add in a well rated large 25mm dia
25k pot in series with the 27k
then you can adjust the pot until the voltages become equal when the amp input
is set at say
5Vrms at 8 ohms.
Then carefully remove the pot without changing the set position, measure what
the adjust
resistance is, and solder in a fixed value of R in series with the 27k and that
should give you equal
output voltages from the phase inverter.
The R12 will always have to be larger than R11 for balance to occur,
so start with them equal, and use a pot to adjust R12 to be larger until balance
occurs.

The values of R11 and R13 are rather low for this sort of circuit.
The bias voltage applied to V2A should be +70V from the anode of V1 EF86.
Therfore the common cathode should be at about +76v, so the DC in R13 18k = 4.2
mA,
so current in each R11 and R12 is about approx 2.1mA each.
So if the B+ = 420V, then you get a 57 V drop across the 27k
so anode voltage = about +363V.
The 355V shown on both anodes on the schematic is fictional, and cannot occur,
since RLs are different.

The combined AC/DC load on the phase inverter is 24k on one side and 27k on the
other as existing
with Rg of 150k and and RLs of 27k and 33k in parallel respectively.
If only 2.1mA flows in each 1/2 of 6SN7, then max +ve swing voltage in 24k is
50.4V
or 35.6Vrms.

To get a greater voltage swing capablity you'd need to
reduce R13 to about say 12k, thus increasing the voltage across each anode RL,
and allowing a higher +ve going swing voltage which may reduce the thd.

But the difference between R12 and R11 increases.
But always R12 will be greater than R11 for balance.

The only way to make the circuit balance perfectly and have R11 = R12
is to use a constant current source between the common cathodes and 0V.

Or else establish a -400V source, and have 106k as R13 instead of the existing
18k,
and still R12 will have to be slightly more than R11, since a slight current
change will still occur
in the 106k.


I hope the operation of the phase inverter / differential / Long Tail Pair
has become easier to understand, but I took awhile to understand exactly how it
worked
and thus what to expect when adjusting it for correct operation.

Patrick Turner.





>
> Dan
>
> "Patrick Turner" <info@xxxxxxxxxxxxxxxxxx> wrote in message
> news:43108D8E.FD8E58F9@xxxxxxxxxxxxxxxxxxxxx
> >
> >
> > Dan Nicotera wrote:
> >
> >> Patrick, thanks for the tips. I'll go over it one more time.
> >> I worked on another pair of these amps about 8 months ago and they
> >> produced
> >> 59-60w RMS before clipping, so I figure these should produce the same.
> >> I'm running them at 65ma as per the original Eico design.
> >> I have checked the input to each power tube and they are the same level,
> >> so
> >> the phase splitter stage seems to be ok.
> >> My plate dissipation is about 490v X .65ma=31.85w.
> >> I'll go and check the neg biases again for each output, but I think they
> >> are
> >> all ok.
> >> Thanks....
> >> Dan
> >
> > 32 watts of idle Pd is above the EL34 rated maximum safe dissipation,
> > and is asking for trouble imho.
> >
> > And you are maybe right about the max power of 50-60 watts perhaps.
> >
> > 60W into 8 ohms = 21.9Vrms, and this would be class AB1 watts.
> > Its a lot, and may perhaps be the absolute max po for the amp.
> >
> > I also plot a graph of max po at clipping onset
> > for various RL, 2,3,4,5,6,8,12,16 ohms when using the 8 ohm outlet.
> > The graph will have a peak at some RL with lower po each side of that RL,
> > but about 60W is all that's possible from EL34 with Ea = 490V.
> > At 60W and a sine wave the Ea will fall considerably since class AB
> > causes the input power from the PS to increase to more than the idle input
> > power.
> > Where you have the tubes biased for maximum class A % of the po,
> > ie, 65 mA, the grid bias applied will not allow a large Vg swing to be
> > applied
> > before grid current is encountered, so in doubt you'd see 60W, ever.
> > If you bais back to say 40mA, the Ea will rise, and max po will increase
> > because
> >
> > more applied grid signal can be applied without being limited by grid
> > current.
> > But still you will see Ea fall at full po since Ia for each tube at full
> > po
> > inceases.
> > With music, the Ea stays about the same since music does not
> > have the same continuous voltage level as a sine wave.
> >
> > Patrick Turner.
> >
> >
> >>
> >>
> >> "Patrick Turner" <info@xxxxxxxxxxxxxxxxxx> wrote in message
> >> news:43108109.CA65E13C@xxxxxxxxxxxxxxxxxxxxx
> >> >
> >> >
> >> > Dan Nicotera wrote:
> >> >
> >> >> I've jsut rebuilt a pair of Eico HF60's and first off I'm only seeing
> >> >> about
> >> >> 50pp output across an 8 ohm
> >> >> dummy load. I've injected in a 1 Khz sine wave and on the scope the
> >> >> lower
> >> >> half of the sine wave seems to be distorting or not quite as linear as
> >> >> the
> >> >> top half.
> >> >> As I incresse the input, the top half of the sine wave starts to clip,
> >> >> but
> >> >> not the bottom.
> >> >> Is only part of the push pull not operating properly? I've tried
> >> >> another
> >> >> EF86, 6sn7 and different EL34's and even 5AR4. The output transformer
> >> >> checks
> >> >> out ok also.
> >> >> Both amps have the same problem.
> >> >> All voltages check out, all components also check out ok.
> >> >> Any ideas guys??
> >> >> Thanks...Dan
> >> >
> >> > The 50V p-p is 17.7Vrms which is an output power of 39 watts.
> >> >
> >> > Isn't that about what you'd expect at the onset of clipping?
> >> > What does the manual say about max power?
> >> >
> >> > Clipping in some tube amps isn't always "symetrical",
> >> > ie, occuring at the same time on +ve and -ve peaks and there are a
> >> > number
> >> > of
> >> > reasons why
> >> > such as the output tubes being not matched, the driver amp topology
> >> > gives more drive to one tube than the other when grid current begins to
> >> > occur.
> >> >
> >> > Have you measured all the signal voltages at eacg anode and grid before
> >> > any
> >> > clipping
> >> > begins to occur, say at ***32 watts?***
> >> >
> >> > I assume your schematic is as at
> >> > http://www.triodeel.com/hf5060s.gif
> >> >
> >> > You should find that the drive voltages to each EL34 grid have the same
> >> > amplitude.
> >> > If they are not equal, then adjust the value of R12 ( which is shown as
> >> > 33k )
> >> > either larger or smaller until the drive voltages are equal.
> >> >
> >> > Each of the grid drive voltages should not be much distorted below
> >> > clipping.
> >> >
> >> > The bias currents at idle can be balanced by adjusting R2, the 10k
> >> > balance
> >> > pot.
> >> > Make sure that without any signal present that voltages across R18 &
> >> > R19
> >> > are equal, and approximately 0.40 volts which means
> >> > about 40mA flows in each EL34 so that anode & screen dissipation is Ea
> >> > x
> >> > Ia
> >> > = 490 x 0.04 = 19.6 watts.
> >> > Check that when the bias is set that the applied -ve voltage to each
> >> > gris
> >> > is
> >> > about the same for each
> >> > tube.
> >> > A difference of more than 10% grid bias may mean one tube is worn, ie,
> >> > needs
> >> > more
> >> > grid bias than the other, so it'll be the other healthy one with less
> >> > grid
> >> > bias
> >> > that clips first.
> >> > Also measure the DC idle voltages across the EL34 grid bias resistors
> >> > R16
> >> > & R19
> >> > of 180k. The voltage should be less than 0.2V, meaning that the grid
> >> > voltage
> >> > is not more than 0.2V *positive* with respect to the negative bias
> >> > voltage
> >> > at the ends of the 10k balance pot.
> >> > Should you read a +ve voltage of say 1V across either 180k of both, you
> >> > have
> >> > either a problem with a leaky
> >> > coupling cap ( C9 & C10 ), or you have a tube/s which are worn out and
> >> > due
> >> > for
> >> > replacement.
> >> > Measure this bias resistor voltage 3 minutes after turn on then agsain
> >> > 15
> >> > minutes later.
> >> > There should be no slow increase of the DC voltage across either bias
> >> > R.
> >> >
> >> > And if you don't like the sound you could try
> >> >
> >> > http://www.geocities.com/dmitrynizh/hf60-dht.htm
> >> >
> >> > But maybe KT88 or KT90 may work better.
> >> >
> >> > Biasing for KT88, KT90, KT66, 6L6, 5881 are all about the same,
> >> > ie, for an equal applied grid voltage the bias current will be similar.
> >> > EL34 is the odd one out, and bias should be adjusted for
> >> > between 35 and 40 mA for EL34 or any of the above tubes if used.
> >> >
> >> > Also make sure that you have the 8ohm load connected to the 8 ohm
> >> > outlet,
> >> > not to the 16 ohm outlet.
> >> >
> >> >
> >> >
> >> >
> >> > Patrick Turner.
> >> >
> >

.

Relevant Pages

  • Re: Can a 211 amp use an 845
    ... > I love the 211 but I was just wondering if the amp can do both. ... heater voltage is the same. ... About 50vrms of grid drive is needed, gain is about 10.2, ... so that the main load seen by the driver is ...
    (rec.audio.tubes)
  • Re: Ill probably get flammed !
    ... to ground 1 ohm R the plate voltage reads around 335v (somewhat less ... volt.The amp was designed for 6V6 finals so I assume that is the reason ... between the plate and grid values for AB fixed bias. ...
    (alt.guitar.amps)
  • Re: NFB101 Crisis Averted
    ... Indeed the the transconductance amp is one with extremely high input ... So there is no input current to the amp, so a tube grid or fet gate ... The current supplied from a voltage source signal flows in R1 and R2 ... If the open loop gain is high ...
    (rec.audio.tubes)
  • Re: Marshall 4140 Amp question
    ... >> running value, which can stress the power supply caps, and any other ... >> This is not quite such a bad thing if the amp employs semiconductor HT ... >> and the resulting unloaded voltage may well come very close to, ... > Charred/burnt 1.5K grid resistors and blown HT fuses. ...
    (sci.electronics.repair)
  • Re: Blackface a Silverface Super Reverb
    ... You'd be better off putting the money into an amp that designed the ... By listening to the tone of various amps and making an informed ... The voltage differences between BF and SF are the bigger issue. ... High voltage on tubes run too cold causes what I believe you are talking about. ...
    (alt.guitar.amps)
Loading