Selecting Capacitor(s) and Resistor(s)

[Adriel]

^^^That!^^^
I'd really recommend trying LTSpice. It's free, and works pretty well simulating circuits, including tube amps, rectifiers etc.

You use it a lot?

 

[Adriel]

Or if you prefer open source - I certainly do - then you can check out Ngspice.
I didn't try it yet, but it's on the list.

I am not that kind of person, I just sit down to a computer and use it, no desire to program or build. Why haven't looked into Linux.

 

[kodabmx]

I will stick to my knowledge XP was is best Microsoft OS. Nothing fancy, gets the job done and don't have to dig for files or programs like Windows 8 and newer (why bought Windows 7 after seeing how terrible 8 is).

8 was no worse than 7... People just didn't like "Metro".

You can hate Macintosh, however, I find it the most logical OS. Amusingly, the first computer I used, back in the 2nd or 3rd grade, was the latest Apple released in about 1993. I remember the green screens and our school getting something called electronic mail, so amazing mail coming on a wire.

Not to me. Unix and derivatives are far more logical to me. Macintosh computers are like cars with automatic transmissions - they require almost no knowledge or skill to operate.
I'm about 10 years older than you I think... I grew up with the Apple 2II in school before they got the original Macintosh. It wasn't until high school that we had real computers. Amiga for multimedia (ahead of it's time) and x86 machines for "business"
I remember it taking about half an hour for a lab of 25 Macs to load Claris works because Appletalk was sh*t.

Speaking of Xerox, their copiers were tough. In the 1980s, Dad was part of a team testing equipment to be supplied to the US Navy. One such test was shaking, basically simulating rough seas conditions. The tests had to be carried out until no further damage could be done or the equipment stopped working. Well, the Navy asked SPAWAR to test a Xerox copier. When it arrived, it was bolted down to the shake table, shake table turned on, and the violence tore the feet and the copier went flying, landing onto the concrete floor top side down, shattering the glass. They turned it back over, one held a piece of paper over the copier about where it be if the glass were there, and the darn thing made a copy that was good enough to use! Strapped it back onto the shake table and ink was leaking out the big bottles. However, the bottles were in place and the door, which held the bottles in, seemed to be holding. However, the high speed photography later revealed the latch was unlatching, the bottles then tip out and spill ink, the door close and latch, securing the bottles back into place. 😆 SPAWAR's recommendation was stronger feet and a stronger latch. Later, a Xerox advert said Government tested.

Dad absolutely loved that kind of work, it quit after Clinton defunded the military, no new equipment to test. He literally became a paper pusher and loathed it, but, he loved us children and so kept working until forced him to retire, couldn't afford the required pay raise (after spending thousands to teach him SolidWorks).

My dad worked for Xerox for over 30 years 🙂 He was involved with IMS and mainframe more than actual hardware though.

 

[jcalvarez]

You use it a lot?

Yes, I do. It's not perfect by any means, but it will give a good idea about what not to do. Also it simulates quite well the most common topologies using popular tubes. The key is finding the right tube model (model=script used by Spice to describe a particulate tube). Most commonly used tubes out there have pretty solid models.

 

[OldHector]

Once you have a supported OS for LTSpice, installation is straightforward, and you can have a simple file with a '.asc' postfix which can contain all the information to simulate a tube circuit. Someone can find that for you easily. Then, if you want to experiment more, there are library files that are easy to add that describe the models for different tubes and transformers. Spice comes with Triode and Pentode symbols, but if you get adventurous you can create or add more.

It is not a substitute for a real circuit, but it is a great sanity checking tool. If I build a new circuit it is nice to simulate it so that theoretical and actual voltages and current s can be compared, dissipation in components checked, and real world changes tested - for instance using a transformer with slightly different characteristics or changing the feedback, or even using another tube for the same job.

 

[kodabmx]

It works on Windows, Linux, and Mac? Unless you're still running DOS, you could use LTspice.
I just prefer an open solution. I'd be surprised if it didn't use .asc or work with the models that already exist - why re-invent the wheel? Even if that heppened, someone would make a converter because it's open sourse and they can figure out how it works and write for it 🙂

 

[jcalvarez]

It works on Windows, Linux, and Mac? Unless you're still running DOS, you could use LTspice.

Seems to be only Windows 64bit and MacOS. No Linux I'm afraid, but maybe it could run under Wine.

 

[kodabmx]

It has run under WINE for years... Not sure about the latest version, but I assume I'd work. Frankly, I don't know why they haven't make a web app for it yet.

 

[jcalvarez]

@Adriel Here is a simulated ANK EL84 Rev F (the schematic you posted). Works pretty well, beautifully delivering 10.5W with a 0.17V peak input signal 🙂.
6BL8 is a close equivalent to the ECF80. 6BL8P is the pentode part, 6BL8T is the triode part. 6BQ5 is equivalent to the EL84.

 

[Lampie519]

The simulations are good to check if no hidden errors have been made. To be sure to simultate “real world” components one should use a Monte Carlo analysis and check if all stays stable.

 

[jcalvarez]

The simulations are good to check if no hidden errors have been made. To be sure to simultate “real world” components one should use a Monte Carlo analysis and check if all stays stable.

Interesting, did not know you can do Monte Carlo analysis with LTSpice. Neat.

 

[Lampie519]

I was using Multisim at the time.
20 years ago i was a Senior Application Engineer at National Instruments and technical trainer. Normally i don’t use any simulations for my designs as i know the results anyway. The models are just not complex enough to let me do the things i do on the breadboard.
OldHector meantioned something similar already.

 

[Adriel]

My dad worked for Xerox for over 30 years 🙂 He was involved with IMS and mainframe more than actual hardware though.

Interesting how we still live in a "small" world.

 

[Adriel]

Yes, I do. It's not perfect by any means, but it will give a good idea about what not to do.

Nothing on this Earth is perfect, it is a Western thought that gave us success and perfection.

Thank you for answering whether you know the program. 🙂

 

[Adriel]

Once you have a supported OS for LTSpice, installation is straightforward, and you can have a simple file with a '.asc' postfix which can contain all the information to simulate a tube circuit. Someone can find that for you easily. Then, if you want to experiment more, there are library files that are easy to add that describe the models for different tubes and transformers.

Can you also load a design into Spice and modify it or use for building another?

Maybe understand the differences in the components or even typologies? Or does the typology need to be specified?

Spice comes with Triode and Pentode symbols, but if you get adventurous you can create or add more.

I have my doubts I get that adventurous.

It is not a substitute for a real circuit, but it is a great sanity checking tool.

Understood.

If I build a new circuit it is nice to simulate it so that theoretical and actual voltages and current s can be compared, dissipation in components checked, and real world changes tested - for instance using a transformer with slightly different characteristics or changing the feedback, or even using another tube for the same job.

I don't have the money or desire to purchase multiple transformers.

Interesting mention feedback, that had previously crossed my mind.

 

[Adriel]

It works on Windows, Linux, and Mac? Unless you're still running DOS, you could use LTspice.

About two weeks ago, I had my SSA interview and the agent was having computer difficulty. Turns out a few weeks prior had switched from M-DOS to an internet based program (with illogical interface). 😳

 

[Adriel]

Seems to be only Windows 64bit and MacOS. No Linux I'm afraid, but maybe it could run under Wine.

There is a Windows XP version, though no longer getting updated. Doesn't matter to me anymore, the XP computer's startup got corrupted and not fixable.

 

[Adriel]

Frankly, I don't know why they haven't make a web app for it yet.

Well, you seem very knowledgeable and enjoy programming, so why not do it yourself? 😛

 

[Adriel]

@Adriel Here is a simulated ANK EL84 Rev F (the schematic you posted). Works pretty well, beautifully delivering 10.5W with a 0.17V peak input signal 🙂.
6BL8 is a close equivalent to the ECF80. 6BL8P is the pentode part, 6BL8T is the triode part. 6BQ5 is equivalent to the EL84.
View attachment 1145692

Did you create this?! 😳

Is this a monoblock? I am not seeing the resistor tying into the other channel. Just curious...

Is there a way to transfer this onto another computer?

I be interested in running it (though understand might have to create it myself) and trying other input voltages, ditto with the four tube (to gain understanding regarding gain).

ECF80 file does not exist?

 

[jcalvarez]

Did you create this?! 😳

Yes, but I did not start from zero, I have many circuits, I got Tubelab's SPP and modified it to match ANK.

s this a monoblock? I am not seeing the resistor tying into the other channel. Just curious...

Monoblock or not does not matter from simulation point of view. All you simulate is the amplifier part, usually one channel, and do not even simulate the power supply.

Is there a way to transfer this onto another computer?

Yes, very easy, you just get the schematic file (*.asc), model files and that's it. Well, provided you have installed LTSpice 🙂

I be interested in running it (though understand might have to create it myself) and trying other input voltages, ditto with the four tube (to gain understanding regarding gain).

I really recommend that, yes. I can send you all tube models I have and schematic files, well, the decent ones.

ECF80 file does not exist?

I could search for a specific ECF80 model, I do not have it, but for all intent and purposes the 6BL8 is the same thing.

I updated the schematic, fixed some bugs and changed the component's references to match the ANK schematic.
Now it looks like this: