My first tube project ! - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 18th August 2012, 01:41 PM   #1
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Default My first tube project !

Hello

Three days ago I finished my first tube project. So, you could consider me as a newbie in tubes. The project is a single ended class A mono power amplifier without feedback. Actually it is a modified version of “Glass House 300Bse” stereo amplifier circuit designed by Andy Groves (Audio Note). In general I used the same circuitry. Major changes I did was the second 300B paralleled with the single of original circuit, the different EI output transformer equipped with a special core of 17,000 Gauss rated at 100W and 1,500Ω primary winding. Also a huge choke of 10H/310mA. At the moment I haven’t possibility to upload the schematic (with voltages marked at all critical points) because my main PC – where the plans are stored – was damaged and is at service now.
So I will make a brief description – though I think many of you have knowledge – for the rest:
Input tube is a simple pentode EF86 – you probably know that simple pentodes in input is the beloved method of Andy Groves in all Audio Note amplifiers – which obtains the full voltage gain of amplifier. It is buffered from the double output 300B tubes thru a 5687 very strong double triode configured as simple voltage follower (0dB gain). The EF86 is AC coupled to 5687 thru a 15nF “mylar in oil” capacitor. The 3rd (suppressor) grid of EF86 it is connected to cathode which is biased by the appropriate 2.2KΩ resistor to GND and bypassed thru a 220μF capacitor. A 470nF “mylar in oil” huge capacitor connected across the 3rd suppressor grid and the 2nd screen grid eliminates the Miller phenomena. The two output tubes are of the brand “Full Music” equipped with solid carbon plates (300B/c). I selected these tubes for their ruggedness compared to mesh plate ones; we need as much as possible power because our speakers are some “heavy”. Plate voltage respect to GND is 450V while the voltage across plate – cathode is 393V as cathode is biased at 57V. The output transformer (made in Greece by G. Antoniadis laboratories - Thessaloniki) proven amazingly good, as under any output load the voltage drop of primary winding is just 7V. It stays cool (not heated) yet after 2 – 3 hours of operation. With an 8Ω non-inductive dummy load connected at output and with a sinus of 0.45Vrms as stimulus in input (that is the input sensitivity or the max signal for full output power) I measured with a DSO connected across the load a peak voltage of 13Vrms, stable, from 50 up to 2,500Hz. That means a maximum output power of 21.12Wrms. Without load the peak output voltage is 19.6Vrms. Later I did the usual bandwidth test according to EIA standards (output set at -3dB of its max value) to locate the -3dB points: The low limit is 12Hz and the high 22KHz, thus bandwidth = 12 – 22,000Hz. For the record, at 10KHz = 8.2Vrms (-0.8dB), at 20KHz = 6.6Vrms (-1.6dB), at 40KHz = 4.2Vrms (-6.6dB), at 60KHz = 2.48Vrms (-11dB) and finally at 100KHz = 0.86Vrms (-20dB).
The power supply includes a 400VA mains transformer. The HT voltage is rectified by a 5C3S (Russian military version of 5U4G and more rugged) and smoothed thru a “Π” filter composed from 3 nice Mundorf M-Lytic HV+ capacitors (those with 4 pins), Allen Bradley carbon composition resistors (can withstand better in high inrush current surge pulses than carbon film) and the huge 10H/310mA choke. Separate power supply is used for the 5687. Filament voltages for all tubes (except rectifier 5C3S) are DC stabilized via three LM1084 – 5A voltage regulators. Mains power consumption after a 30’ preheat time is stabilized at 1.2Aac. Output noise, though the whole device is still “bare” (cover plates and transformer covers are still in process) and so non screened, varies between 2mV to 4mV. I hope will drop down to 0.5mV after the shielding of everything.
What impressed me from my first contact with a tube amplifier (I am a solid state guy for 35 years) is that: a tube amplifier actually does not clip the output signal be it so is overdriven. Being sure that its max output voltage capability is 13Vrms/8Ω I tried bigger sinus stimulus in input than the max 0.45Vrms. Of course it is obvious that the sinus signal starts to distort presenting a rounding and a small tilt in the “falling” edge of sinus but does not presents a hard clipping like in solid state amplifiers (the signal is “latched” at the level of supply rails and presents a straight “duty cycle” like in square waves) a thing that is very dangerous for the speakers. Instead this, the tube amplifier simply “rounds” the corners of a clipped signal. Output transformer also plays a role on this.
For audition tests, please don’t ask me. I like very much music of any kind (from rock up to classic, jazz, folk and anything sounds nice in my ears) and I hear every day enough hours. But to be focused on a sound device it should produce significant amounts of distortion! For that reason I entrust only what I can measure with my instruments and I study thoroughly the technology of electronic parts from active and passive components up to relays, switches and cables. For example, in this project I had contact with carbon composition resistors for first time in my life. Also with “Mylar in Oil” foil capacitors, tantalum resistors, those M-Lytic HV+ of Mundorf, UPOCC Silver wire etc, etc.
To the present I attach 4 screen shots I taken from my DSO that show the behavior of amplifier in square waves at the 4 most common frequencies: 100Hz, 1KHz, 10KHz and 20KHz. Also a sinus that shows the peak voltage amplitude. In all measurements a 8Ω dummy load was connected at output. When I receive back my workstation PC I will make and the FFT analysis (how much is the 2nd harmonic?) and I will upload all schematics. Pictures of device in few days when chassis will be completed.
I will appreciate very much any comment and any suggestion. Thank you very much.
Attached Images
File Type: jpg 100Hz.jpg (124.9 KB, 931 views)
File Type: jpg 1KHz.jpg (129.1 KB, 897 views)
File Type: jpg 10KHz.jpg (126.2 KB, 876 views)
File Type: jpg 20KHz.jpg (129.4 KB, 851 views)
File Type: jpg PeakPower8.jpg (130.2 KB, 828 views)
__________________
Best Regards FOTIS ANAGNOSTOU

Last edited by fotios; 18th August 2012 at 01:46 PM.
  Reply With Quote
Old 19th August 2012, 09:48 AM   #2
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Default Schematics 1

Hello

Here are attached the schematics of Power Supplies. The HV marked is for the high voltages while the LV marked is for the tube filaments. Also the schematic of amplifier. All schematics are presented in artistic mode but the value and the identity of each device is the true, as well all voltage markings. All measurements have taken with the mains voltage stabilized at 228Vac thru an autotransformer which is continuously callibrated by a microcontroller to the preset value, thus is not a simple variac. Because the poor resolution of pictures i also attach these in pdf format. You can open the pdf files to see most clearly the schematics.
Thank you.
Attached Images
File Type: jpg 300B-PSE_HV.jpg (217.5 KB, 358 views)
File Type: jpg 300B-PSE_LV.jpg (126.5 KB, 352 views)
File Type: jpg 300B-PSE_AMP.jpg (133.1 KB, 283 views)
Attached Files
File Type: pdf 300B-PSE_HV.pdf (37.9 KB, 45 views)
File Type: pdf 300B-PSE_LV.pdf (28.1 KB, 29 views)
File Type: pdf 300B-PSE_AMP.pdf (41.2 KB, 77 views)
__________________
Best Regards FOTIS ANAGNOSTOU
  Reply With Quote
Old 19th August 2012, 01:34 PM   #3
diyAudio Member
 
Join Date: Nov 2011
Location: Europe
Hi,

congratulations for finishing your project!
Schematics are a little hard to read, but i think i understand the amp part. Well, the way G2 and G3 of EF86 are connected is new to me and i would like to understand it better. In my builds it is connected "normally" as pentode with some nfb for lower distortion (it sounds a little edgy without it), or as triode without nfb.
If there is enough gain you could try some nfb around EF86, it might sound better this way.

Cheers
  Reply With Quote
Old 19th August 2012, 05:43 PM   #4
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Quote:
Originally Posted by GlowSignal View Post
Hi,

congratulations for finishing your project!
Schematics are a little hard to read, but i think i understand the amp part. Well, the way G2 and G3 of EF86 are connected is new to me and i would like to understand it better. In my builds it is connected "normally" as pentode with some nfb for lower distortion (it sounds a little edgy without it), or as triode without nfb.
If there is enough gain you could try some nfb around EF86, it might sound better this way.

Cheers
Thank you so much for your kind words and your comments GlowSignal !
You are right about the unusual connection of G2 and G3, it is this huge 0.47μF "Mylar in Oil" capacitor connected between them. I have seen in Philips datasheet the usual way of connecting G2 to Gnd thru a 0.05μF cap.
EF86 works as a voltage gain stage. Looking at the Philips datasheet and taking into account the given values of anode - cathode - G2 resistors and the 217V anode voltage, this tube should work with Av = 170. With the given input sensitivity of 0.45Vrms, Av = 76Vrms that exceeds the max output of 44Vrms given by Philips. So the required gain is fulfilled. BTW EF86 has much higher gain than any triode but needs some special precaution due to its sensitivity.
In other products of Audio Note, like the kits "Legend monoblock" and "Interstage monoblock", Andy Groves again makes use of a pentode in input, this time a 6SH7. In the "Legend" the 6SH7 drives directly the two output 300B tubes and again Andy has connected a 0.47μF cap between G2 and G3. In the "Interstage" the 6SH7 drives the two 300B output tubes thru an interstage transformer and in this design Andy does not use this 0.47μF cap accross G2 and G3.
From my solid state experience i know very well that each voltage gain stage presents large ammounts of Miller capacitance which is translated in overshoot and oscillations in the collector (anode for tubes) of VAS transistor. To eliminate it, we connect a "Miller compensation" capacitor of some pF (usually 15 to 100pF) accross collector and base. I think that Andy makes the same by inserting the 0.47μF cap accross G2 and G3, while in the "Interstage monoblock" Miller capacitance is eliminated by the primary winding (is a good coil) of interstage transformer so the cap accross G2 and G3 is absent. I am not sure for the above described, but in my project i tried to remove the 0.47μF cap and i had big oscillations from 500Hz and above. For that reason exactly i make the tests with square waves that i presented in my first post. You can see that all square waves are clean without any overshoot.
By sure the 0.47μF cap offers a form of feedback. I don't know what kind of feedback you mean exactly. Please advise me, and GREAT THANKS AGAIN.
__________________
Best Regards FOTIS ANAGNOSTOU

Last edited by fotios; 19th August 2012 at 05:51 PM.
  Reply With Quote
Old 19th August 2012, 06:50 PM   #5
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Default Square Waves

For 30 years, i allways use square waves to test any of my solid state projects. The same i did and in my first tube project, you can see the results in the screen shots of first post. The use of a square wave as stimulus in input, it can reveals many usefull informations about the circuitry of project. Ringing, overshoot and oscillations show a bad compensation of Miller capacitance of the Voltage Gain Stage. It also shows a great instabillity of amplifier that is mandatory, especially in single ended without feedback tube designs. Negative feedback offers some form of protection against instability and its absence is a great risk for the circuit. A sinus wave used as stimulus can not reveals these oscillations or overshoot, if exist. I have seen in some tube threads screen shots from oscilloscope that show these ringings. For that reason we use a compensation foil capacitor of some nF arround input tube, either connected to Gnd or accross G2 (screen grid) and G3 (suppressor grid). I think that this capacitor offers some form of feedback from mid to high frequencies in input tubes, especially in pentodes which offer very high voltage gain so are very sensitive in external RFI, EMI etc. infections.
Thank you for your attention.
__________________
Best Regards FOTIS ANAGNOSTOU
  Reply With Quote
Old 19th August 2012, 07:49 PM   #6
ChrisA is offline ChrisA  United States
diyAudio Member
 
Join Date: Jan 2008
Quote:
Originally Posted by fotios View Post
Hello

Here are attached the schematics of Power Supplies. The HV marked is for the high voltages while the LV marked is for the tube filaments. Also the schematic of amplifier. All schematics are presented in artistic mode but the value and the identity of each device is the true,...

These "artistic" schematics are VERY hard to follow. Wires are crossed up inside the very small tubes symbols.

I looked at the scope shots. I'm betting a few percent harmonic distortion in the single digit range. This is I think the goal of both these SET amps as well as guitar amps - to add a bit of distortion. How much is a mater of taste.
Looking forward to the FFT numbers
  Reply With Quote
Old 19th August 2012, 09:20 PM   #7
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Quote:
Originally Posted by ChrisA View Post
These "artistic" schematics are VERY hard to follow. Wires are crossed up inside the very small tubes symbols.

I looked at the scope shots. I'm betting a few percent harmonic distortion in the single digit range. This is I think the goal of both these SET amps as well as guitar amps - to add a bit of distortion. How much is a mater of taste.
Looking forward to the FFT numbers
Great thanks Chris for your comments! Are so precious to me, though i have enough experience with electronics design that is my first tube project and i feel some stressed, nothing more.
I ask your forgiveness for the "artistic" schematics, only these i saved in my old PC (i use this at the moment) before my good PC was fried! It needs new power supply and motherboard, i will have it ready in 3 days, the original and clear schematics are stored on it and i will upload these immediatelly. Also i will make the FFT analysis.
I appreciate very much your incisive look, that is what i expected, a man who can estimate the performance of a circuit from the DSO screen shots of square waves! It is obvious that you are a true audio electronics engineer - designer probably working in one of these famous Hi-End audio USA companies! Thank you so much!
For the record now, from the 10KHz and 20KHz waves shape it is obvious that the EF86 is somehow overcompensated. For that reason the -3dB upper frequency limit lies at 22KHz. If i used a smaller cap than the 470nF monster, say 220nF, i think that i could obtained a -3dB upper frequency limit at 40KHz but... in the cost of increased distortion. I know the concept behind tube amplifiers, is the high level of 2nd harmonic (mosfets also obtain a signifficant ammount of it compared to "poor" BJTs) that sounds nicelly in the ears of people. So i try to do a compromise between conflicting things. I don't like e.g. a 10% THD yet at full power! I could accepted a 5 to 6% THD as much. The amplifier is addressed for a home Hi-Fi music system, not for electric guitar, i know from these things as i was working enough years in live stages. The lead guitar performer likes a... ton of distortion at 130dBSPL in bass frequencies where the thing is technically difficult to obtained. It is mostly a mater of speakers, not so much of the tube head!
Another one issue is the output power. These 300B/c are amazingly strong. I tried them with 500V on plates, and i got almost 15Vrms accross the 8Ω dummy load just before clipping occurs and that means an output power of 28 clear American Wrms (i allways use the EIA methods and standards). OTOH, i really dither to push so hard the output 300B/c, i still don't have experience of their service life, just that are stronger compared to mesh plate ones.
Thank you again Chris.
__________________
Best Regards FOTIS ANAGNOSTOU

Last edited by fotios; 19th August 2012 at 09:46 PM.
  Reply With Quote
Old 19th August 2012, 10:14 PM   #8
ChrisA is offline ChrisA  United States
diyAudio Member
 
Join Date: Jan 2008
Quote:
Originally Posted by fotios View Post
... a man who can estimate the performance of a circuit from the DSO screen shots of square waves! It is obvious that you are a true audio electronics engineer - designer probably working in one of these famous Hi-End audio USA companies!...
No, nothing like that. I have an old Tek, 365 analog scope and an older Heathkit distortion analyzer that has an analog meter that reads out in THD. No computers or FFT inside it's all done with vacuum tubes. This simple meter is useless with modern solid state amps as they all read "zero" THD but works well for tube amps with the NFB loop open. I think the meter is a valuable tool for breadbroarding vacuum tube based audio circuits and checking components.

If you see a sine wave on the scope and have an analog THD meter you can quickly learn to judge THD percent based on the shape of the sine wave. I see blunted tops on the waves but not really bad ones.

But the best use of a THD meter is the "residuals" output. The meter can subtract out the fundamental and give the remaining signal. Then you put this on a scope and you can see only the distortion. It is easy to learn to see the cause by looking.

You can do all of this on a computer too but I hate to connect a computer to 400 volt tubes. If I do that enough times I'll cross up some wires
  Reply With Quote
Old 23rd August 2012, 08:45 PM   #9
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Default Clear schematics & valid voltages

My workstasion PC is again in order repaired and upgraded with new AeroCool PS, Asus M/B, Intel i3 3.4G CPU and 8GB RAM! Though the expense was big (340 Euro) for the difficult times we pass in Greece i finally am happy. The worst is that the PC damage (fried Power Supply and M/B) caused from bad luck, i did a measurement in the output of power resistor after the cathode of rectifier 5C3S (480V) and the probe of multimeter was sliding on chassis causing a short circuit. My PC was in operation and fried . Probably the 480V HT passed thru the mains earth conductor to PC. BTW this 5C3S proved a nonster! With its solid carbon plates it doesn't give a damn about any short-circuit. The same the MILLS 50W cathode resistor!
Well, as i promised i attach the actual schematics with valid voltage marked on each point. Tommorow i will make the FFT analysis of amplifier and i will post the results.
Fotis
Attached Images
File Type: png 300B-PSE 0_AMP.png (133.5 KB, 283 views)
File Type: png 300B-PSE 0_HV.png (53.4 KB, 230 views)
File Type: png 300B-PSE 0_LV.png (45.1 KB, 140 views)
__________________
Best Regards FOTIS ANAGNOSTOU

Last edited by fotios; 23rd August 2012 at 09:05 PM.
  Reply With Quote
Old 23rd August 2012, 09:15 PM   #10
fotios is offline fotios  Greece
diyAudio Member
 
fotios's Avatar
 
Join Date: Feb 2007
Location: Δραμα - North Greece
Default some first pictures

Though the amplifier is still not completed, i attach two general view pictures before the wiring and one that shows a detail of HT power supply pcb stacked with long spacers on the 5687 tube power supply pcb.
Attached Images
File Type: jpg amp-gen5.jpg (178.8 KB, 292 views)
File Type: jpg amp-gen7.jpg (455.0 KB, 281 views)
File Type: jpg p.s.1.jpg (376.9 KB, 231 views)
__________________
Best Regards FOTIS ANAGNOSTOU
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
possible project for me, tube preamp. theChris Tubes / Valves 20 31st October 2012 12:47 PM
My OLD Project....... cheap tube audio project!!! tube-lover Tubes / Valves 4 8th May 2008 09:26 PM
First tube project. dome406 Tubes / Valves 15 28th September 2005 10:12 AM
first tube project Spooky1 Tubes / Valves 4 12th August 2005 02:23 PM
My first tube project Brannigan Tubes / Valves 42 7th November 2003 09:47 AM


New To Site? Need Help?

All times are GMT. The time now is 01:06 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2