• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

el84 amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Re: Experimenting with EL84 PP

kooltubes said:
I am wondering if someone can suggest a good circuit ( no NFB), yet simple! I don't need much power; 5 watts are more than enough.
I have got a few EL84 and Ecc99 laying around, plus 2 pairs of pushpull output tx with ultralinear taps.

By the way, I've bought (very very cheap) a pair of brand new output transformers ( 9K, ultralinear, taps at 20%) and need to know what can they be used for?
Thanks for your input :)
You have all the ingredients. ECC99 differential front end to an EL84 trioded output stage thru the 9k OPT would be fine.
 
kooltubes said:
I was thinking biasing the ECC99 (differential) at about 18 ma as suggested by JJ, common cathode via LM317T and 68R resistor between Adj and out connected to -16V
I just mormally use a fet based CCS. Don't like the 317, even removed them from use as CCS in shuntregs
Output stage triode connected shared 270 ohms cathode resistor.
Have I done my homework right?
Thanks
More like 130R for a shared cathode resistor at a B+ of 250V and Ia of 35mA/tube.
 
Ex-Moderator
Joined 2003
kooltubes said:
I was thinking biasing the ECC99 (differential) at about 18 ma as suggested by JJ, common cathode via LM317T and 68R resistor between Adj and out connected to -16V

You definitely don't want to use 317 in the audio department, it's a voltage reference plus a 741 connected to a sluggish power transistor. If you're only passing 18mA, you can make something much better out of discrete transistors. A cascode would do nicely. Transistor type isn't critical, neither are the LEDs.
 

Attachments

  • 18ma ccs.gif
    18ma ccs.gif
    8.4 KB · Views: 508
You definitely don't want to use 317 in the audio department, it's a voltage reference plus a 741 connected to a sluggish power transistor. If you're only passing 18mA, you can make something much better out of discrete transistors. A cascode would do nicely. Transistor type isn't critical, neither are the LEDs.

Hi EC8010,

Do I just connect it to both cathode of ECC99?
If a simple circuit like yours could make wonders, why not?
 
Ex-Moderator
Joined 2003
Traditionally, you choose the anode load resistors and anode current so that the anode voltage is half the supply voltage. When you know a little more, you draw loadlines on anode characteristics and carefully choose the operating point. When you know a bit more than that, you look at your available stock of resistors and mutter, "I expect that'll be alright." Really posh people surround the circuit with test gear and finely adjust the value of the emitter resistor in the CCS for optimum performance. Sometimes, this causes a noticeable difference.;)
 
Hi EC8010,

I am not rushing anything, therefore I've got a ruler, a pencil and trying to draw different loadlines.
Got myself some books, a scope (don't know how to use it as yet :eek: ), a multimeter; enough to keep me busy this winter!

I am trying to learn something....:) I don't want to depart from this world like an idiot who only knew to buy parts and solder them!

The good thing about this forum is that I get the stimulation I need, hoping it'll pay one day!

Thanks
 
Wow...in the last three days I've lost some good answers

kooltubes, for me it's the same...I am trying to learn something about this beautiful world. A friend of mine can lend to me his scope ( he buy it this summer but he never start to learn how to use it...) and I'm very interested to learn how use it for my hobby. Have someone a good link about this argument?
Next saturday I'll go to Pordenone just to see one of the most interesting electronic fair in the north Italy...I'll try to find some good books about this argument ( and for sure something about tubes and fet's design)... two ecc99 and an el34's quartet.
 
EC8010 said:

If you're only passing 18mA, you can make something much better out of discrete transistors. A cascode would do nicely. Transistor type isn't critical, neither are the LEDs.
I have some questions. Today I've read in a book which are the mathematical models of a BJT. This interesting start to understand how this circuit works:
http://www.webalice.it/jlc891/page7.html

The driver stage is an hibrid srpp that use a particular led that turn on with only 1ma...well we can talk about this late.
Diego Nardi uses this design in several projects because he can obtain full gain, low distortion, very low output inpedence, similar to one catofollower, and high slew rate.
I'm trying to understand this circuit but maybe you can teach me something about it :D
I want to try something like it to load the driver stage of my el 34se amp ( it use a sigle e88cc) but first I want to understand how it works.
why to use a led? Which parameters have I to consider for the design of similar stages? How does It works? And which are the differences between catode ccs and this hybrid srpp?

Thanks...sorry again for my english and for this long post :angel:

Mark
 
Ex-Moderator
Joined 2003
The Nardi circuit works in a very similar way to the cascode circuit that I posted, but it's a little easier to understand the cascode first, then look at the Nardi circuit.

Looking at the left hand side of the cascode circuit, there are two LEDs in series with a resistor. There is 16V across the whole chain. A typical red LED drops about 1.65V at 10mA, so let's drive 10mA through them. That means that the LEDs drop 3.3V total, so there must be 16V - 3.3V = 12.7V across the resistor. If there's 12.7V across the resistor and 10mA through it, then its value is 12.7/0.01 = 1270 Ohms. 1k3 is near enough. So far, we've simply lit a couple of LEDs, but the significant thing is that LEDs make quite good voltage references, so in addition to producing some light, we also have a voltage reference at the bottom of 1.65V.

Now let's look at the lower transistor. To make life easier to understand, we'll treat the -16V rail as if it were at 0V. (You'll soon see why.) It takes about 0.65V across the base/emitter junction of a transistor to make the collector/emitter path conduct. Let's assume that the transistor is conducting. If we have 1.65V on the base, and the base/emitter junction is dropping 0.65V, then there must be 1V on the emitter. Neither the base/emitter drop of the transistor or the LED voltage change very much with current, so that 1V is pretty well fixed.

We now add a resistor between the emitter and the -16V rail. It has 1V across it, and we know that we would like our CCS to pass 18mA.

Small transistors like the BC549C have a very large current gain. This current gain is either known as hfe (hybrid model, forward mode, emitter as common terminal), or beta. The current gain is the ratio of collector current to base current. For the BC549C, the manufacturer guarantees that beta > 400. This is very significant. It means that if 18mA is flowing into the collector, only 45uA has to flow into the base. Our reference chain (the LEDs and 1k3 resistor) passes 10mA, so drawing 45uA away from it will not change the voltage across the lower LED.

Take the last two paragraphs together, and we see that if we put a resistor between the emitter and -16V rail, it will have 1V across it., almost no matter what. So let's do that. 1V/0.018A = 55.55 Ohms. So why did I write 62 Ohms? Must have been a mistake!

Now for the clever bit. If we take the collector of the lower transistor to the cathode of a valve, it will force 18mA to flow, and the voltage at the cathode of the valve will adjust itself to accomodate as the valve wears out or if it is swapped for a different valve.

In practice, as the voltage between the collector and emitter of a transistor changes, so does the current. But only slightly. The way we get around this is to observe that looking down into the collector, the current gain of the transistor has a multiplying effect of the emitter resistor. In our case, we used a 56 Ohm resistor, and the current gain is 400, so looking down into the collector, we see 400 x 56 = 22400 Ohms. An ideal CCS would have infinite resistance.

Let's add another transistor on top of the first (making it a cascode). If the upper transistor also has a current gain of 400, and the resistance looking down to the -16V rail is 22400 Ohms, then the resistance looking down into the upper collector is 22400 x 400 = 8960000 Ohms. 9M is a pretty good constant current sink (unless you're Gary Pimm).

Finally, to make the lower transistor work, we need to drop some voltage from the collector to its emitter, and that's why the second LED is there, it forces 1.65V across the lower transistor.

The Nardi circuit works in a very similar way, but is very slightly compromised by passing the current from its reference chain into the load.
 
EC8010 said:
The Nardi circuit works in a very similar way to the cascode circuit that I posted, but it's a little easier to understand the cascode first, then look at the Nardi circuit.

Ehmmm....I love you :joker:
I'm printing your answer just to read it with attention, thankyou very much.

EC8010 said:
The Nardi circuit works in a very similar way, but is very slightly compromised by passing the current from its reference chain into the load.

Maybe this is the reason why he uses this led and has wrote, in his articles, that only with this one the circuit sound well

http://www.fairchildsemi.com/ds/HL/HLMP-K150.pdf

Again thankyou for your answer :wave:
 
Ex-Moderator
Joined 2003
You're welcome. Yes, I expect that is exactly why Nardi recommends that particular LED. In case you're wondering why he should do something less than perfect, the alternative would be to take a resistor down to 0V. The trouble is that it would dissipate a lot of power, making it hot and expensive.
 
Thanks EC8010

Hi EC8010,

Thanks for the CSS circuit, I've just finished building the amp; it sounds good for a 5 watts'ish (PP EL84 triode connected).

It worked fine, the first time, without any hum and bang.
I had to tweak the current on the ECC99 to get a better sound; I am running the differential driver at about 21 ma pe tube, yielding to an anode voltage of approx 140 volts.

I am still learning.....oops...:eek: I haven't opened the scope box, as yet, but that my assignment for next week.
Cheers!:)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.