Tube-Town Forum
Technik => Tech-Talk TT-Projekte => Thema gestartet von: Mjork am 12.09.2021 12:37
-
Hello TubeTown friends,
I am new here and Dutch therefore I hope English will work,,,,
I did learn something about electronics but it has been a while, now I an rediscovering all that.
I have built the Emma kit and found the output was noisy and undefined.
I checked all the DC voltages, grounds and of course the elements, but all seemed fine.
Then I started using a scope to trace a 110Hz signal from my looper through the circuits.
I found out it worked all fine incl the signal going into the grid of the ECC99 (pins 2 and 7).
Strangely enough the signal from ECC99 to the push-pull transformer (pins 1 and 6) looked different.
The frequency was doubled. the amlitude weakened a bit at the same time.
When checking the speaker out put I indeed saw 215Hz coming out while sending 107Hz in.
Does anyone recognize this issue maybe?
What is the next thing to check here?
-
I think also the signals coming from pins 1 and 6 should be in counter phase?.
On checking the push-pull Hammond 125C that the resistance between B2 (DC feed) and signal leads in not the same. red-blue is 128ohm and red-brown is 159ohm. I would expect equal resistance, as the windings should be the same.
Ich denke auch, dass die Signale von den Pins 1 und 6 in Gegenphase sein sollten.
Bei der Überprüfung der Gegentakt-Hammond 125C, dass der Widerstand zwischen B2 (DC-Einspeisung) und Signalleitung nicht gleich ist. Rot-Blau ist 128 Ohm und Rot-Braun ist 159 Ohm. Ich würde einen gleichen Widerstand erwarten, da die Wicklungen gleich sein sollten.
-
I think also the signals coming from pins 1 and 6 should be in counter phase?.
On checking the push-pull Hammond 125C that the resistance between B2 (DC feed) and signal leads in not the same. red-blue is 128ohm and red-brown is 159ohm. I would expect equal resistance, as the windings should be the same.
Ich denke auch, dass die Signale von den Pins 1 und 6 in Gegenphase sein sollten.
Bei der Überprüfung der Gegentakt-Hammond 125C, dass der Widerstand zwischen B2 (DC-Einspeisung) und Signalleitung nicht gleich ist. Rot-Blau ist 128 Ohm und Rot-Braun ist 159 Ohm. Ich würde einen gleichen Widerstand erwarten, da die Wicklungen gleich sein sollten.
Hi Mjork
the doubling of the frequencies is part of the behavior of the kathodyne PI when the powertubes start to draw grid current. I'm a bit astonished, that this even happens with a ECC99, but with 6V6s it is perfectly normal >> 5E3 Tweed Deluxe.
regards
Jochen
-
Hi Jochen, thank you for picking this up.
OK, I must say I am a bit new with all this.
But putting in 107Hz and getting out 215Hz surely could not be the desired effect? It gives you a different sound I would think?
I started my investigations on the Emma when after building it I had noticed:
- I do get sound when playing a string but only when putting the volume over 50%
- when I play more strings or when I hit one string harder the output becomes really scrambled, like the speaker is ripped
- when I connect one of the two ECC99 grids inlets to ground sound comes out as I would expect.
So surely something must be not OK as I do not get the proper sound, but what?
Even, when I connect one of the two ECC99 grids inlets to ground, the output signal on BOTH output pins doubles when I use the scope to observe.
See attached pics. It not only sound better, it also looks better ;)
-
So what does happen ther?
I would expect loss of output signal when connecting on of the inlet signals to ground, but here I gain output instead?
-
Hi,
I started my investigations on the Emma when after building it I had noticed:
- I do get sound when playing a string but only when putting the volume over 50%
- when I play more strings or when I hit one string harder the output becomes really scrambled, like the speaker is ripped
- when I connect one of the two ECC99 grids inlets to ground sound comes out as I would expect.
As I wrote already earlier I guess that you have a problem with the ground somewhere. I know you said that you checked everything twice and everything is connected correctly but it is not otherwise the amp would work as it should.
Now you have a scope so use it to troubleshoot the amp. Check the first stage, is it okay then the next etc. If you do not understand how the amp, the PI or something else works then you must learn the basics first otherwise you can not make sure that the amp is build correctly and works safe and stable.
Easy but helpful: enter the circuit in Spice a simulate it and compare the result of the simulation with your measurements. This could help a lot.
Best regards, Dirk
-
If you provide us with some high-resolution images of the wiring inside the amp we may be able to trace out what is wrong with it. :topjob:
Apart from that I agree with Dirk.
-
Goodmorning,
I did indeed check the other circuits in the amp, see attached log file.
I have no list of values that could be expected but the AC signal seems to nicely increment, what do you think?
Will try that Spice simulation software when I have some more time, good tip.
Attached the wiring pics. The component connections like R15 to R16 are made on the back side, not touching the housing/ground.
-
The signal on the grids is absolut 400mVpp. Far away from any overdrive Kondition.
But the plate-signals are heavenly distorted.
I also see no gain.
Try to check the heater current & voltage on this tube
If this OK than the ECC99 maybe faulty.
-
No, no, the grid on that picture is much higher: at 1 (V/division) x 2 (divisions) x 10 (pen setting) = 20VAC. And close to 0 VDC
Only when I close the volume pot from 50% to approx. 20% then finally the plate signal looks like a perfect sine again.
At that point the grid voltage is 9VAC and the plate voltage is 1,5 VAC.
Heater voltage: 6.6V between pins 4/5 and pin 9.
-
Hi,
I just run Emma thru the simulator with these settings:
All potentiometers a 50%
Input Signal 100 mV @ 1 kHz Input Chanel 2 Low
8 Ohm Load
Curves
red: Testpoint Grid V2B
blue and green: Testpoint C8-R17-R20 and C9-R18-R21
Maybe this helps.
Best regards, Dirk
-
Hello TubeTown friends,
I am new here and Dutch therefore I hope English will work,,,,
I did learn something about electronics but it has been a while, now I an rediscovering all that.
I have built the Emma kit and found the output was noisy and undefined.
I checked all the DC voltages, grounds and of course the elements, but all seemed fine.
Then I started using a scope to trace a 110Hz signal from my looper through the circuits.
I found out it worked all fine incl the signal going into the grid of the ECC99 (pins 2 and 7).
Strangely enough the signal from ECC99 to the push-pull transformer (pins 1 and 6) looked different.
The frequency was doubled. the amlitude weakened a bit at the same time.
When checking the speaker out put I indeed saw 215Hz coming out while sending 107Hz in.
Does anyone recognize this issue maybe?
What is the next thing to check here?
the doubling effect can happen in a push pull OT when it is wired incorrectly,the 2 halves are in phase,i don't know if you have access to the center tap junction to try reversing one side
-
Dirk, thinks for taking the efftort of simulation.
I was starting with this, nbut it is more complicated then I thoought.
--> Where can I find the .sub files or the data therein for these: 12AY7, 12AX7 and ECC99?
I changed my input to 1kHz at 0,1 V and measured all points again, see attachment.
The 3 curves you got from LTspice are spot on in my amp.
It is the ECC99 where things seem to go wrong.
What i first thought to be doubled frequency is actually a dent pulled in the positive wave of the output sine which makes it look like doubled.
When i measure it I see frequency peaks at 1000 and 2000Hz.
If I turn the vol pot back the dent disappears and at about 25% it is close to a sine, see pics. At all volumes the amplification is negative.
-
You can use a 12AU7 instead of a ECC99 for testing. Maybe you got one.
The PI is working correctly ? This is what you should first check.
Google knows where you can get the models for the tubes. I also posted a sim a few days ago including some models and a potentiometer lib which you need.
Here it is: https://www.tube-town.de/ttforum/index.php/topic,24782.0.html
Best regards, Dirk
-
I managed to get my hands on a new ECC99 but the result is the same, so the tubes all seem OK.
The PI works good I think, it shows two perfect sine waves in counterphase as output, 15VAC.
The input is a 15,5VAC sine. Voltages as in Dirks simulation. The DC voltagesare as per the drawing.
So that leaves the ECC99 outpput signal still to be strange.
Grid has perfect sine 15VAC as input, the output plate voltages is no longer a perfect sine and around 11VAC.
DC voltages are close to 260VDC (plate) and 8.5VDC (cathode).
What more can I check?
And what can I check on the push-pulll trafo?
I noticed there is a difference in the resistance between the two inut sides (red to blue = 124 Ohm, red to brown is 160 Ohm), is that relevant? The number of windings must be the same on both sides.
-
The difference in the resistance is OK. I see the same on my trafo.
Next point to check is the heater. Both filaments working?
Try to check the current in the heater circuit. Voltage readings are not helpful in this case.
-
OK, this is something I did not try yet.
If I disconnect the filament wiring from all tubes en make a loop over just the ECC99 pins 4+5 and pin 9 I get 6.6VAC between the pins and the current through the MM shows 5.1mA flow through the circuit. Both the heaters are glowing in the tube. When taking the 12AY7 I see approx the same values
Is that a correct value?
-
R23 and R24 are both connected to ground ? Both pins in the tube socket are not damaged and the tube pins are having a good connection with the socket. The value of both resistors is correct ? Checked it with a DMM ?
I would say that one system of the tube is not working correctly and I would check all connections and components which are connected with the ECC99 first.
If you swap the blue and brown wire of the OT does the misformed half wave stays at the same position or does it move to the other half of the sine wave ?
Best regards, Dirk
-
Oh, I though one of the pictures which you posted of the scope shows both plates together but it doesn't. It would be good to see both plates together too.
Best regards, Dirk
-
R23 and 24 are connected to ground: yes
Tube and pins look OK, i have a second ECC99 and the curves did not change
all resistors I did check with a DMM indeed
Have not tried swapping blue and brown yet.
The pic showing both plates together is attached
-
?...through the MM shows 5.1mA flow through the circuit.
MM no idea
Anyway, the heaterloop runs at 300mA + 300mA + 800mA + 30mA ~ 1430mA. 30 mA is for the two 100Ohm resitors.
Don't know your skills, so let me Show up an easier way.
Disconnect the wires from pin 9
Disconnect the bridge betwenn pin 4 and 5
Now test the resitance betwenn pin 9 and 4. Should be less than 5 Ohm for a cold tube
Do the same with pin 9 and 5. Should also be less than 5 Ohm.
If you see these values, the connection from the Socket to the tube is OK. Current can flow.
-
MM = DMM
I also expected higher currents. spec sheet gives 300mA at 6.6V also.
Verified with a loose power supply, 10VAC, and got approx the same values.
It must be something with the DMM settings. Drives me mad but I cannot find it.
My other DMM does not measure current at AC so cannot check.
Anyway, the resistance I checked. Pins 9 to 4 and 9 to 5 are both around 2.6 Ohm when really cold,
-
What DMM Model you use?
For the next step to solve that Problem we need two 10k resistor. 1watt or more would be nice.
-
Do you have a connection between the two grids ?
You have tested an other ECC99 with the same result so the tube should not be the problem.
I guess if you de-solder the wire which connect s grid 1 OR grid 2 of the ECC99 the amp works (of course only with half power). You should do this too to test both sides of the output transformer
If both wires are connected the amp does not work because both systems of the ECC99 are "on" at the same time. This is what the scope shows (if the measurement is correct).
The PI seams to work correctly too so I could imagine that you have a connection between the two grids which causes that both systems of the ECC99 are working in parallel.
Best regards, Dirk
-
When I swap two Wires on the primary side of output trafo, the plate signal shows similar behavior as your amp.
-
I have swapped the blue and the brown wire (the input sides of the trafo, but the result is the same.
Disconnecting one side of the trafo shows as per attached pic. Both waves are identical sine shape and the plate is 40VAC with the grid at 15VAC.
So now there is amplification in that final stage and the dent in the wave is gone.
-
For an 8 ohm speaker you use 4 and 6 ?
-
Disconnecting one side of the trafo shows as per attached pic.
And the other side ? How does it look when you disconnect the second plate and connect the first again ? It should look the same just mirrored. Does it ?
Did you check if you have a connection somewhere between the two grids of the ECC99 ? Maybe somewhere on the board or at the socket ?
Best regards, Dirk
PS: I have also some PCBs of the Emma (not online yet) which is more failsave to use. If you are not able to find the mistake to use the PCB is another option because you do not have to make any onboard connections. But you should try to improve the soldering point and short the wires regardless what you use.
-
Ähmm
may throw the possibility of "grid current" back into the discussion?
Let me guess - the output of the cathodyne which is showing the spike is the anode connection...
Kirchoff tells us, that the signal at the two working resistors of the cathodyne have to behave equal, under the assumption that there is no current leaving the circuit.
Jochen
PS: try to use PSPICE Tube Models which simulate grid current condition. In the PSPICE SEction of the forum - look for the latest models of Thaddäeus... They might provide this feature. I have not tried them yet, but i am confident.
-
@ Gunter: 8 ohm speaker, connection 4 and 6 indeed
@Dirk: other side looks almost the same, not mirrored, but bit flattened though, see pic. But that does not really mean anything as it depends from what point the signal is picked up, sometimes I see the signal just flip while measuring.
So to chech if they are mirrored I disconnected both plates from the trafo. Then looked at the signal on the plate (induced signal?) Then you can see they are opposed to eachother, see other pic.
Thanks for offering the PCB but I will first lift the board from the housing again and check/redo the soldering of the 6 resistors and interconnections. Something for the coming friday.
@ cca88: the signals after the PI are nice sines and 100% mirrored from eachother.
-
Today I have soldered 6 new resistors on a loose print and connected those in instead of the ones on the onboard ones. Connected between C8/C9 and the ECC99. Checked all values again throughout the amp.
All end up the same, output signal also. So the onboard resistors are fine.
All has been checked now and is correct all the way up to the ECC99 inputs and also the ECC99 is OK. What
remains is the output transformer it seems to me.
When one side is disconnected I get amplification from 10 to 40 VAC, when I connect the other side also the output signal collapses from 40 to less then 10 VAC.
As can be seen on attached pictures, the output signals are in counterphase (as they should be), but this is only seen when I close the volume pot to around 5%.
Then I go up to 20% and you seen the dent in both wave kicking in on the pos side.
Then when going up further to 50% the signals look like two signals that are in phase.
So why could cause this to happen? :(
Anyone ideas? Thanks, Marc
-
to see if the transformer is the problem, you can use two resistors.
Anything from 1k to 10k will be fine for testing.
it will only change the plate voltage and maybe some clipping will show up.
grid signal should be less than 12V for nice curve. Above the signal shows some distortion but will stay out of phase.
-
Anyone ideas?
You have a sineweave generator and a scope so you can check the OT. To get some ideas please see: http://www.marstran.com/Measuring%20Primary%20Impedance.htm
How did you connect the OT ? All pins not connected are "open" and not connected to ground or something else ?
Best regards, Dirk
-
@ Gunter, I tested the system with two 10k resistors instead of the OT. Good suggestion!
Signal from the plates is now a nice sine and they are exactly in counter phase, like a mirror.
You are correct, when going up in volume over 10VAC on the grid the outsput signal starts to flatten on the top.
But no dent is coming down and the signal phase stays mirrored.
Does this invalidate the OT now? Then I will order a new one straight away.
N.B. The resistance I measure in the OT are 160 and 130 Ohm per coil, way lower then the 10kOhm installed now.
@Dirkj: I did not get to do the OT tests yet, but indeed all 4 other pins are just left open.
-
a coil has two resistors. once the dc resistor, in this case 130/160 ohms. and then there is a second resistor in a row. this
is called impedance.
this changes with the frequency. the 125c has an impedance with an 8 ohm speaker of about 13kohm at 1000 Hz.
Looks like that there is somthing wrong with the output trafo.
-
@Dirk: I do not have a function generator, but I used a 1kHz signal I got from youtube and got in the aux input of my looper.
This works well and the looper volume knob increases the output voltage.
If I follow correctly the example in http://www.marstran.com/Measuring%20Primary%20Impedance.htm
these are the results
- when connecting the looper to the OT input wires blue and brown I cannot get more than 580mVAC (RMS, measured with DMM). Red wire is left open.
- input 580mVAC, output between pins 4 and 6: 3 mVAC
- this gives a VTR of 193
- Imp ratio then is 37249
- multiplied by 6 (pin number) gives 223494. According to the example this then is the Primary Impedance of the OT?
What do you conclude from this?
-
Hi,
if you have an other output transformer then give it a try. I have never seen such a kind of failure but it looks like that the source of the problem is the OT.
Best regards, Dirk
-
No I did not have one so I ordered one from some obscure company called Tube Town?
New OT arrived and performing the test from http://www.marstran.com/Measuring%20Primary%20Impedance.htm
gave really different results:
- when connecting the looper to the OT input wires blue and brown the output did not drop like with the old one, so I could keep the 1 VAC.
- input 1000mVAC, output between pins 4 and 6: 24 mVAC
- this gives a VTR of 42 (old one 193)
- Imp ratio then is 1764 (old one 37249)
- multiplied by 6 (pin number) gives 10584 as a primary Impedance (old one 233k)
So I had good hopes on going to use this new OT, and indeed now the Emma works as to be expected!
17VAC on the input of the ECC99 gives 95VAC out.
Both sides perfect sine in counterphase and signal to speaker is 4,8VAC.
Seems like I can pick up the gitar again instead of scope and soldering iron and finally get to experience that so called tube sound!
Pfffffft.... It has been a long ride. But with a steep learning curve which you do not get when first time right!
Thank you to the supporting members!
-
:guitar: :topjob:
-
That's good but still a little bit strange. I guess there is an internal connection between the two windings or something like this. Never seen before and I would really like to know what it is. Anyway, please contact us via the contact form for a refund.
Best regards, Dirk
-
Meine Vermutung ist, das das Eisenpaket durch irgend einen Umstand magnetisiert worden ist. Dadurch geht der Trafo schon Bei kleinen Signalen in eine Art Sättigung.
-
@Dirk, do you want me to send the OT back, so you can check it out?
-
@Dirk, do you want me to send the OT back, so you can check it out?
no, this will be a waste of money.
Best regards, Dirk