| lumanauw |
Hi all,
Wanted to ask the tube expert about this schematic. I have problem with PS, so I wanted the tube section works in lower voltage.
How much lower the tube can go? I wanted the tube to be able to work in +/-2V p-p sinusoidal without defect or clipping. Will it go to +/-15V for the tube? |
|
|
| fdegrove |
Hi,
| quote: | | I wanted the tube to be able to work in +/-2V p-p sinusoidal without defect or clipping. Will it go to +/-15V for the tube? |
The tube will probably still handle the +/- 2V p2p but what you'll probably be feeding it is quite likely 2V RMS or put differently 6 V p2p.
With an effective B+ of 30V I'd expect some distortion with the 6DJ8/ECC88.
What you can try is use a 6GM8/ECC86 instead or otherwise explain us what the problem is you're experiencing with the GC.
Cheers,;) |
|
|
| lumanauw |
Hi, fdgrove,| quote: | | The tube will probably still handle the +/- 2V p2p but what you'll probably be feeding it is quite likely 2V RMS or put differently 6 V p2p. |
Yes, I wanted to take all kind of input without problem. Maybe there are CD player / preamp that have much output.| quote: | | With an effective B+ of 30V I'd expect some distortion with the 6DJ8/ECC88. |
The datasheet of ECC88 have max voltage included, but not min voltage. I read somewhere that tubes cannot work in order with too low voltage. Just how much low voltage can the ECC88 works properly in this design? Is +/-35V is minimum, or it can go lower?| quote: | | What you can try is use a 6GM8/ECC86 instead or otherwise explain us what the problem is you're experiencing with the GC. |
All I can get here is ECC88/6922. Also, I read somewhere that ECC86 does not sound so good.
Why I ask the question? I wanted to built this design into 2 possibilities. First the tubed gainclone. It is ok with the supply, since it is the same value.
Second to built a tubed preamp (replacing the gainclone with opamp). In this tubed preamp, the opamp needs only +/-15V, so to make things simple, could the tube take the same +/-15V also (or I will need other supply just for the tubes) |
|
|
| fdegrove |
Hi,
| quote: | | Just how much low voltage can the ECC88 works properly in this design? Is +/-35V is minimum, or it can go lower? |
If it were me I'd go higher but not lower.
Have you considered using a voltage doubler as a powersupply?
It's not the best there is but short of adding another xformer it might save you a lot of hassle if your current xformer can handle it.
Also, have you put your question to mr Rasmussen himself?
Cheers,;) |
|
|
| lumanauw |
Hi, fdegrove,
I will ask a question. But I take your answer is Higher, instead of Lower.
I'm not familiar with tubes. In this cathode follower, is the input signal have the same height as the output signal (as buffer) or it will have somekind of magnification? |
|
|
| fdegrove |
Hi,
| quote: | | But I take your answer is Higher, instead of Lower. |
Higher than the +/- 15 VDC rail you suggest, yes.
I noticed you put the question to mr Rasmussen already.
What you need to work with tubes is an understanding of loadlines.
It's a little beyond the needs of what you're doing now but if you plan to work with tubes in the future you'll find it quite interesting.
Hope it was of help,;) |
|
|
| moamps |
Hi,
my last experiments with ecc88 show good behavior with power supply of +/-19V, FET cs with 2mA and Va ca 30V. (grid is about -10V). Maximum AC voltage thru buffer was ca 5Vpp.
Standard ECC88 buffer circuit (grid on 0VDC) can't work good on +/-15V PS, IMHO.
Regards |
|
|
| fscarpa58 |
HI| quote: | | Also, I read somewhere that ECC86 does not sound so good. |
At low voltage ECC86 sounds very good. In my op. better than
ecc88, e88cc,6922, pcc88 , 6dj8 etc.
| quote: | | Standard ECC88 buffer circuit (grid on 0VDC) can't work good on +/-15V PS, IMHO. |
I do not agree, these tubes work and sound rather good at low
voltage. The following cir, for instance, shows high dist.
at 2Vrms ( 0.5% about) but , irrespective to this, it sounds good.
Bye Federico |
|
|
| lumanauw |
Hi, Moamps,| quote: | | my last experiments with ecc88 show good behavior with power supply of +/-19V, FET cs with 2mA and Va ca 30V. (grid is about -10V). Maximum AC voltage thru buffer was ca 5Vpp. |
Do you have drawings on your schematic? I'm not quite understand it. The supply is +30V and -10V?
Hi, Federico,
I'm also confused with your drawing. The supply is floating and where is the output is taken.
What does the THD of 0.5% behave, if we use big rating power amp? Will it produce hiss, or distorted sound, or something else, or nothing? |
|
|
| moamps |
| quote: | Originally posted by lumanauw
[B]
Do you have drawings on your schematic? I'm not quite understand it. The supply is +30V and -10V?
|
Hi,
schematic looks similar to second schematic on this site:
http://www.diyzone.net/diy/referenc...dj8buffer_2.htm.
Cathode is on ca -10V potential. Supply is +/-19V.
Regards |
|
|
| moamps |
| quote: | Originally posted by fscarpa58
I do not agree, these tubes work and sound rather good at low
voltage. The following cir, for instance, shows high dist.
at 2Vrms ( 0.5% about) but , irrespective to this, it sounds good.
|
Hi,
in your design Va is ca 25V what is similar to my (30V).
I was referring to Joe's buffer schematic in +/-15V PS condition with grid tied to 0V where buffer can't work satisfactory, IMHO.
Regards |
|
|
| fscarpa58 |
hi
| quote: | | in your design Va is ca 25V what is similar to my (30V). |
not quite. With the model I use I obtain about
15-20 V. as Va
see for instance the following
ciao |
|
|
| fscarpa58 |
hi
| quote: | | Will it produce hiss, or distorted sound, or something else, or nothing? |
It gives rich and warm sound.
Federico |
|
|
| Tyimo |
Hi!
What about this, with current mirror?
Would you use AC or DC voltage for your heather?
Tyimo |
|
|
| Tyimo |
Next one...
I think Mr. Ernő Borbély is a great expert of the low voltage tube amps. He wrote the ECC86 can easily work on+/-24V and even better with CCS! The ECC88 could be good too, but with a bit higher voltage :34-40V
Tyimo |
|
|
| fscarpa58 |
HI
I experimented with both Ecc86 and Ecc88 from
10 V up to 20 V between plate and cathode in a CF
configuration.
ECC86 always shows half the distortion of Ecc88 in
these conditions.
Federico |
|
|
| fdegrove |
Hi,
Ciao Federico,
| quote: | | ECC86 always shows half the distortion of Ecc88 in |
That's to be expected, the ECC86/6GM8 was made for this kind of working voltages, the ECC88/6DJ8 needs at least twice as much to perform comparably well.
Cheers,;) |
|
|
| lumanauw |
Hi, Tyimo,
Nice idea. What is the sound of your suggested CCT (using current mirror).
In the cct in post#1, what happens if we change the 10K with CCS, about 3.5mA? Will this CCT still works? If it still works, what is the merit/drawback? |
|
|
| Tyimo |
Hi Lumanauw!
Sorry, but the idea is not mine! I will build it in 2 weeks and afther I can tell you about the reasults.
Greets!
Moamps!
How I have to setup and adjust the bias in your your design with CCS? What I have to measure? ca. 2 mA?
Thanks |
|
|
| moamps |
| quote: | Originally posted by Tyimo
How I have to setup and adjust the bias in your your design with CCS? What I have to measure? ca. 2 mA?
|
Hi,
I use fet 2SK30 with IDSS current (gate is connected to source) witch is ca 2mA. In posted schematic current is set to some higher value, IMHO and can be measure indirectly by measuring voltage on known value of R3 (ohm law). It is possible to lower this current with trimmer in FET source for some degree.
Regards |
|
|
| lumanauw |
I'm not familiar with tubes/ECC88.
What is the bias current for ECC88 to get best audio result?
Transistor can be cascoded, like the example below. Mr. Borbely even cascode his fets down as 2V.
If I run the cct at about 3mA, how low can ECC88 be cascoded (Vplate-cathode) and still working properly? |
|
|
| fdegrove |
Hi,
| quote: | | What is the bias current for ECC88 to get best audio result? |
That will depend on your personal taste but I'd go for 10-15mA.
| quote: | | If I run the cct at about 3mA, how low can ECC88 be cascoded (Vplate-cathode) and still working properly? |
By cascoded do you mean with a fet on top of the triode?
Cheers,;) |
|
|
| lumanauw |
| Sorry, the pics of cascode example |
|
|
| lumanauw |
Hi, Fdegroove,
| quote: | | That will depend on your personal taste but I'd go for 10-15mA. |
Waw, thats big. I first tought only about 2-3mA.
| quote: | | By cascoded do you mean with a fet on top of the triode? |
Yes, the the tube will be a follower only (like pic in the first message), but the 10k resistor will be replaced with CCS, and plate will be cascoded with transistor. In transistor I hear better performance with cascode. Is it applicable also in tubes? |
|
|
| fdegrove |
Hi,
| quote: | | Is it applicable also in tubes? |
Sure.
However, if I understand you correctly you intend to replace the 10K cathode resistor from post # 1, the IGC with a tube buffered input, with a constant current source as in Moamps circuit a bit higher up.
So far so good.
However, I don't understand what you want to do with a fet on top of the triode.
Won't you end up with a stack of CCSs in the tail of the transistor?
Instead of doing something like that, if it would work at all, wouldn't it be better to improve the working conditions of the CCS in the tail of the triode by adding a regulator_ provided you have spare voltage to drop_in front of it?
Making a cascode out of a triode and a transistor could probably be done but I just don't see what you could possibly do with it here.
The buffer circuit is not supposed to have any gain, it's there to drive the chip with current, not voltage.
Maybe I just misunderstood,;) |
|
|
| lumanauw |
Hi, Fdegrove,
I will explain why I want to put cascode in tubes.
I learn in transistor, that distortions comes from fluctuating current and fluctuating voltage. The remedy for fluctuating current is to make bias current as big as possible, so the signal current is small compared to bias current. This will make the current induced distortion smaller.
It is the same as fluctuation voltage problem. They way to eliminate distortion from fluctuating voltage is to fix the voltage of the transistor to a certain point, with cascoding. The picture of cascode that I submit is not right, the 100N and 3V9 zener should be tied to emitors of BC560, instead to rails. This will reduce the fluctuating-voltage distortions, since the BC560 will always works in (3V9-0,6V=3V3). So no matter what signal there is, the BC560 will always have steady 3V3 Vce voltage. I tried this and hear the result more detailed, so I want to put steady Vplate-cathode in tubes, maybe they will result the same as transistor.
Also, in transistor, by putting low Vce by cascoding, we can give bigger bias current, since the dissipation will be lower due to low Vce.
My problem is how low can this ECC88 be cascoded? I'm afraid if too low it will ruin the performance. Will ECC88 cascoded to 10V plate-cathode with 10mA bias? |
|
|
| fdegrove |
Hi,
| quote: | | I will explain why I want to put cascode in tubes. |
I understand what you want to do but if you look at the circuit Moamps posted you'll see that the tube is current regulated already.
The reason it sits at the cathode end of the tube is that if you'd put it on the plate side it would replace a plate load resistor.
In a cathode follower you can't use one or it wouldn't be a cathode follower anymore but an anode follower.
By doing that the circuit would perform in a totally different way, it would amplify the voltage at the input and it would be taken from the plate (anode), something we don't want here as we strive to have a low output impedance for the buffer circuit.
If you look at Moamps circuit you will see a MosFet configured as a constant current source set by varying VR 2.
The picture is in post # 16.
Imagine it would operate form a regulated -24 VDC supply and it would be as close as you could come to an infinite load impedance supply for the tube.
This is exactly what you want to achieve with what you describe, only you should look at the circuit sort of upside down if you like as there are no P-channel tubes.
The cathode follower shown here is a little unusual as it uses a bipolar supply of +/- 24 V.
To the tube and the AC current flow in the circuit it is as if the supply is a 48 V B+, it would work exactly the same way, the only reason that it is used here is that the chip is using a low voltage (compared to what is common for a tube circuit) taken from a bi-polar supply already.
| quote: | | My problem is how low can this ECC88 be cascoded? I'm afraid if too low it will ruin the performance. Will ECC88 cascoded to 10V plate-cathode with 10mA bias? |
So, no, from the above that won't work as the cascode as you call it is already in place as per Moamps design.
Cheers,;) |
|
|
| lumanauw |
Hi, Fdegrove,
Thanks for the explenation. You right, it has no sense to cascode the plate if the tube is cathode follower. My mistake.
But your explenation gives me another question.
What makes a tube do not function properly in low voltages? Is it the voltage or the current that is not sufficient?
If it is the current, maybe tubes can work in low voltages, with big current forced. (Again, this is another maybe).
Take example of post#16 cct. It is cathode follower with +/-24VDC rails and CCS in the bottom.
What happens if this circuit is operated in +/-15V, but the CCS is forced tobe 10mA? Will this CCT still works?
Another question. If the current through the tube is quite big (like 10mA compared to 1-2mA) will the tube have shorter life? |
|
|
| fdegrove |
Hi,
| quote: | | What makes a tube do not function properly in low voltages? Is it the voltage or the current that is not sufficient? |
It's just that most tubes were not designed with low operating voltages in mind, you'll go outside their linear operating point which will increase distortion levels.
The ECC88 can work with lowish voltages but don't expect large voltage swings under those conditions.
Contrary to most transistors, tubes are made for voltage amplification, starve them on voltage and they'll compress the dynamic range...something they're used for in studios BTW.
| quote: | | Take example of post#16 cct. It is cathode follower with +/-24VDC rails and CCS in the bottom. |
Look at a cathodefollower as an impedance transformation device like a buffer (which it is), it doesn't need to swing volts so B+ can be rather low depending on the choice of tube.
+/- 24 V still is seen by the device as a total of 48 VDC.
| quote: | | What happens if this circuit is operated in +/-15V, but the CCS is forced tobe 10mA? Will this CCT still works? |
It will still work but probably also distort too much.
Operation points are chosen from the loadlines on the curves of the tube.
The curves are often, but not systematically, published together with other characteristics of the tube such as heater voltage, transconductance and amplification factor among others.
Quite often the usable range is relatively large but there are limits.
| quote: | | Another question. If the current through the tube is quite big (like 10mA compared to 1-2mA) will the tube have shorter life? |
No. Provided the maximum dissipation is not surpassed life will not be shortened.
Mostly too high a plate voltage kills the tube first so lower voltage is sometimes advisable.
Naturally you need to consider what you expect the circuit to do for you when you need to decide on operating parameters such as class of operation, desired output power, etc.
Cheers,;) |
|
|
|
