It would seem the Akai/Roberts Schematic has components that are bang on the Philips datsheets
220k plate 2.2k cathode and 1m screen...am I wrong?
220k plate 2.2k cathode and 1m screen...am I wrong?
Close but not exact. The bottom table on sheet 2 with a B+ of 200V calls for RA=220K vs your last schematic showing 250K, Rk 2.2K vs your 2K. This brings your anode voltage down a bit.
If you are stuck with B+ of 170V, this may be the best you can do.
The 100K in the drain of the FET should be decreased for better drive, especially if you are driving a 10K input (Most common for SS equipment). You will have to watch the dissipation of the LND150 as it is limited to 0.74W at 25C. TO-92 is 160degree C/W thermal resistance.
Dropping it to 33K would give an ID of about 1.5mA for a dissipation of about 0.19W for a temp rise of 30C.
What kind of enclosure is this going to be in?
If you are stuck with B+ of 170V, this may be the best you can do.
The 100K in the drain of the FET should be decreased for better drive, especially if you are driving a 10K input (Most common for SS equipment). You will have to watch the dissipation of the LND150 as it is limited to 0.74W at 25C. TO-92 is 160degree C/W thermal resistance.
Dropping it to 33K would give an ID of about 1.5mA for a dissipation of about 0.19W for a temp rise of 30C.
What kind of enclosure is this going to be in?
HI I realize it's not exactly the same as the schematic.
With the LND150 follower I'm seeing plenty of gain into my rig. I don't need any more gain. Are you saying the LND150 will run cooler with a lower value resistor to ground on the source? For hasty trying I put another 100K in paralell , making 50K I didn't notice a difference in gain, but if it runs cooler/safer, why not? It's in the original roberts akai enclosure. THe follower circuit is built onto the bottom of the board, mostly on an attached 3 point lug strip
With the LND150 follower I'm seeing plenty of gain into my rig. I don't need any more gain. Are you saying the LND150 will run cooler with a lower value resistor to ground on the source? For hasty trying I put another 100K in paralell , making 50K I didn't notice a difference in gain, but if it runs cooler/safer, why not? It's in the original roberts akai enclosure. THe follower circuit is built onto the bottom of the board, mostly on an attached 3 point lug strip
It gets hotter with a smaller resistor, but when the load impedance is low, it can handle larger signals with a smaller source resistor.
It will run hotter, but have lower distortion when driving a low impedance load.
With RS=100K, it will dissipate 0.063W and run roughly 10C above ambient.
With RS=50K, it will dissipate about 0.126W and run about 20C above ambient.
With RS=33K, it will dissipate about 0.19W and run about 30C above ambient.
If there is ventilation in the enclosure and airflow around the parts, there should be no issue with going down to 33K, possibly lower.
With RS=100K, it will dissipate 0.063W and run roughly 10C above ambient.
With RS=50K, it will dissipate about 0.126W and run about 20C above ambient.
With RS=33K, it will dissipate about 0.19W and run about 30C above ambient.
If there is ventilation in the enclosure and airflow around the parts, there should be no issue with going down to 33K, possibly lower.
Don't worry, still alive and kicking.
Currently trying to recover data from my crashed notebook.
FWIW I was scammed by EASEUS , so now trying to use alternative recovery software.
Guess within the weekend will have a couple free hours, reread the thread to refocus and hand draw something.
Take care.
Currently trying to recover data from my crashed notebook.
FWIW I was scammed by EASEUS , so now trying to use alternative recovery software.
Guess within the weekend will have a couple free hours, reread the thread to refocus and hand draw something.
Take care.
Designers often use the datasheet examples, for good reason.
"Old school and boring" if you wish, but reliable.... they should know about their own stuff 😉
As of your earlier question, glad you solved it, will answer anyway because it might help others.
Also to confirm I run faster than the bear 🤣🤣🤣
So starting from the beginning:.
Ok, let's start with your original preamp.
Good in its intended use, it falls apart trying to drive a very low impedance.
It probably expected a 1M next grid load, worst case 500k ... you are loading it with 10k
For a "best engineering practice" design, we must consider both "small signal" performance and "large signal" one.
Many forget considering the second, then "gremlins appear" unexpected.
Your original schematic was:
It is fed about 250VDC, can put out 50VRMS into a proper load (say 1M), but if driving 10k there is a "hidden in plain sight" attenuator: 250k:10k
Horrible impedance matching, huge 25:1 attenuation, very real possibility of damaging a high input impedance preamp your friend might connect there someday, the works.
Uploading this for now, will continue on possible solutions later.
Definitely better than no buffering at all.
That said, it's still not the best to drive such a low impedance load
I mentioned before considering "small signal" and "large signal" for a more realistic approach, also possible output signal possible Vpp swing.
Here you added a high voltage "tube substitute", it would have been the same using a triode.
Output impedance is way lower because of cathode/source follower 100% NFB ... but that applies at low signal levels only.
If/when you need higher signal out, truth is you still have a 100k resistor in series with output.
To be more precise, positive signal half can supply "as much current as needed" because it comes from the active device, BUT negative half is pulled down by the 100k cathode/source resistor and nothing else.
The hidden elephant in the room is that if that preamp is asked higher output it will grossly distort in a very unsymmetric way.
Current source and sink capabilities are completely different.
For low level signals we have a (NFB created) low source impedance, but for high level, 100k is there.
You may imagine a diode in parallel with 100k, pointing towards the capacitor-load : low impedance one way, 100k the other.
I posted this which to me seems very clear and self evident, but if you want "the drawing", so be it.
In a way, "same as Hoffman suggestion" ... BUT .... using a simple cheap FET instead of a semi-exotic LND150 or even an IRF820.
Fed from 25V supply, its source resistor a more useful 4k7 or 10k (considering we just left Tube Land and are interfacing with standard SS stuff)
Signal level will be up to 5VRMS instead of former 50VRMS (per EF86 datasheet suggestion), load driving is way better too.
Still unsymmetrical because of its geometry but way better than before, and more realistic.
Here is the Op Amp version, and it will truly "drive anything" in Preamp Land.
Bonus points: low source impedance all the time , low or high signal conditions, and always symmetrical.
Boy I truly love Op Amps, been using them since 1970, starting with pioneer uA709 and continuing with the king of the hill for a long time: RC741
As they say, the rest is History 😄
SS and LS mean small and large signal.
As you see, I added or changed nothing, just drew what was said on my earlier post.
Hope this helps.
The first drawing on post #49 was poorly edited and includes part of an earlier drawing on top, please disregard it.
Still on a phone which does not let me edit posts 🤷
Still on a phone which does not let me edit posts 🤷
HI thanks for all the info and drawings. I have the skill to build things but not the knowledge to design them (yet) So while your original explanation is likely clear, for me the implementation isn't. That's why this helps and is much appreciated. I might try the FET and op amp routes (though I will have to conjure up a lower voltage power supply one way or another) for the next units I reno
When you say 'you may imagine' are you suggesting a diode in paralell with the 100K resistor will be helpful? And if so what sort of diode would be useful in parallel with the 100K? I have some 18V zeeners, some iN4007s. And when you say pointing towards the capacitor load, , you mean negative to ground and positive towards the capacitor? Thanks again
Thanks.
No, on the contrary, the built_in/"hidden" diode explains the terrible assymetry of cathode followers when poorly designed/used, in this case if trying to drive a very low impedance.
I am well aware of it, because it is used on purpose 😲 in Guitar amplifiers to create much loved assymetrical distortion in iconic amplifiers such as Fender Tweed, Marshall, and their countless clones.
But then those stages are customarily driven to rail_to_ground clipping and beyond.
Hi Fi designers never abuse them that much, so only a few are aware of limitations.
Of course, if you stay within low-ish signal levels, say a couple Volts, any of those works.
A similar example: Op Amps show very low output impedance, sometimes a couple ohm or less..... does this mean they can drive Speakers?
No way!
They are typically limited to, say, 5 mA output or not much more.
Same thing.
No, on the contrary, the built_in/"hidden" diode explains the terrible assymetry of cathode followers when poorly designed/used, in this case if trying to drive a very low impedance.
I am well aware of it, because it is used on purpose 😲 in Guitar amplifiers to create much loved assymetrical distortion in iconic amplifiers such as Fender Tweed, Marshall, and their countless clones.
But then those stages are customarily driven to rail_to_ground clipping and beyond.
Hi Fi designers never abuse them that much, so only a few are aware of limitations.
Of course, if you stay within low-ish signal levels, say a couple Volts, any of those works.
A similar example: Op Amps show very low output impedance, sometimes a couple ohm or less..... does this mean they can drive Speakers?
No way!
They are typically limited to, say, 5 mA output or not much more.
Same thing.
Okay I understand what you were saying now. That's too bad as the follower circuit does provide a lot more gain, but when it does clip the syne wave is assymetric.
Whilst I can build and design many different things from scratch electronic circuitry is not one of them. I have built myself several very nice sounding guitar amps that I use live etc. (I tend to want VERY clean amps so I lean towards fender designs with some mixing and matching, my last one was a stereo deluxe reverb that I mostly use..wish I'd used the ampeg super echo twin reverb design on that. Very practical, clever and nice sounding)
However my approach is pretty well by route I build them like one would build a model. I know how to troubleshoot existing proven designs, but not how to troubleshoot a circuit as to it's effectiveness. Time for some more self education I guess.
Is there any way to tweak the LND150 follower circuit to make it more effective/versatile? Somebody suggestive a smaller size resistor to ground , 50k or 33k
Whilst I can build and design many different things from scratch electronic circuitry is not one of them. I have built myself several very nice sounding guitar amps that I use live etc. (I tend to want VERY clean amps so I lean towards fender designs with some mixing and matching, my last one was a stereo deluxe reverb that I mostly use..wish I'd used the ampeg super echo twin reverb design on that. Very practical, clever and nice sounding)
However my approach is pretty well by route I build them like one would build a model. I know how to troubleshoot existing proven designs, but not how to troubleshoot a circuit as to it's effectiveness. Time for some more self education I guess.
Is there any way to tweak the LND150 follower circuit to make it more effective/versatile? Somebody suggestive a smaller size resistor to ground , 50k or 33k
You can always use a very low value, even 10k, but then do not feed it 250VDC but, say, 30-40V or so ... which you still pull from the 250V rail you already have, no extra supply needed, only a voltage dropping resistor and an extra filter cap.
This source follower can't be direct coupled to EF86 plate but needs a biasing voltage divider.
Just read this 5 times until it clicks 🙂, don't ask me for detailed drawings 😄
This source follower can't be direct coupled to EF86 plate but needs a biasing voltage divider.
Just read this 5 times until it clicks 🙂, don't ask me for detailed drawings 😄
Thanks again. I understand to some extant what you are saying, but I really don't know how to figure out what sort of voltage divider a LND150 needs to bias it. You mean a 1M/100k as mentioned above?
What cheap FETs would you recommend?
Would you still have a cap between the EF86 plate and the LND150. some have said not needed while my electrical engineer brother in law thinks it's a prudent idea
What cheap FETs would you recommend?
Would you still have a cap between the EF86 plate and the LND150. some have said not needed while my electrical engineer brother in law thinks it's a prudent idea