Hello friends! I want to tell you about the creation of my own tube ULF on the 6P36S. This is a push-pull with a fixed offset according to S. Sergeev’s scheme.
It was during the summer holidays, there was a lot of time, there was knowledge, there was desire. So I decided to make some kind of tube stereo amplifier for listening to music. Yes, yes, it was the tube one, I never liked “stones”.
I started thinking about what circuit to take, with what lamps? I decided to choose something from the lamps I had. For some reason I didn’t want to use 6P15P and 6P43P lamps. I was considering the 6P13S option, but read a negative review. In the end I settled on 6P36S. I also liked her appearance: so big and impressive. And the reviews about it are positive.
Tube amplifier circuit by S. Sergeev
High power and the presence of suitable tubes in the driver stage persuaded me to build my first push-pull.
The amplifier sounds very good and has the following parameters:
Supply voltage 360 Volts
Power consumption from the network 170 W
Sensitivity 0.7 Volt
Output power 40 W
Armed with a list of parts, I went to the market. Unfortunately, this is the only place where you can buy radio parts in my city.
Power transformer TSA-270
As a supply transformer, I wanted to use an unwinded standard transformer, for example TS-180. Only TCA-270 was found for me, which I bought and went home, quite satisfied. Although the winding is aluminum (index A), it provides the passport parameters.
I connected TCA-270 like this:
Conclusions 7-17 used to obtain the bias voltage of one channel
Conclusions 4-14 for anode supply of one channel
Conclusions 11-(21-12)-22 for 6P36S filament of one channel. Terminals 20-(10-10’)-20’ to 6N1P incandescence of both channels.
The windings I have given provide exactly the required voltage and current.
Output transformer TS-180
I used a TS-180 as an output transformer and connected it according to the circuit proposed by Manakov.
Surface mounting board
I would like to draw your attention to my non-standard way of creating boards. The fact is that I don’t etch the signet, I cut the foil on the PCB into squares measuring 2 cm x 1 cm and come up with a diagram for wiring the parts. For example:
The diagram according to which I soldered the parts onto the board.
The lines between the squares are jumpers. There are also longer connections, they are drawn as lines. The arrows on the squares indicate the place to which the wires from the lamp panels, or power, etc. will be soldered. For example, 3L1 means that the 3rd pin of the first lamp (6N1P) is soldered there, 8L2 means that the 8th pin of the 2nd lamp (6P36S) is soldered there. Those. the first digit is the pin number, the letter is simply divided, and the second digit is the lamp number.
I'm not suggesting you do the same, it's just that this method is simple and saves time and money. It turns out to be a “canopy” on PCB.
Here's the board I got
First start
The first launch was successful. There is only one pity: at that time I did not have a multimeter at hand. The old one is broken, I haven't bought a new one yet. In general, I set it up using a barbaric method: I adjusted the offset so that it was minimal, but the anodes did not turn red.Everything seems to be working, it’s time to hide it in the case. And then I was very lucky: an acquaintance of my father gave me an old non-working guitar ULF “Minor”, which was immediately gutted.
The case from “Minor” fit perfectly, two TS180 output transformers (lying) and 2 boards of the ULF itself fit into it, but the TCA-270 power transformer did not fit. Therefore, a separate case was made for the power supply from chipboard and plywood. The top of the body is covered with self-adhesive tape.
Holes for the panels were drilled. Everything is installed, and now the long-awaited launch in the case. I turn it on and... I almost went deaf from the whistle as the lamps warmed up! Cursing, I turn off the power and go to the market to buy a shielded wire, buying it with a reserve.
Fighting arousal - shielding
To eliminate self-excitation, which manifests itself as a whistle, it was necessary to shield the grid circuits of the 6N1P triode and the wires in the grid circuits of two 6P36S. The 6P36S anodes were not shielded. The whistling stopped.I start it up again and hear: one channel works great, but in the other there are clicks instead of bass, plus a sluggish and hoarse midrange. Of course I was upset at this point. And what have I not tried? the reason turned out to be incorrect mesh screening. The screens were not grounded correctly.
I redid it, grounded the ends of the screens with a star - this is when all the wires go to one point and to the body. The interference and excitement disappeared, the clicks in the bass stopped, the bass became clear, elastic, and both channels began to sing equally beautifully.
Details
In the amplifier I used MLT and OMLT resistors with a power of 1-2 Watts. I bought all resistors with a power of 1-2 Watts. Capacitors C3 and C4 are type K73-17, and C2 was supplied by BMT-2. I have a 6N1P light bulb in the driver, which can be replaced with a 6N23P without any modifications to the circuit. But I like the sound of the 6N1P better.The 6P36S output tubes, to increase the amplifier power, can be replaced with 6P42S, setting a higher quiescent current, somewhere around 106 mA (drop across the cathode resistors 1.06 V)
Fixed offset setting and feature
After purchasing the multimeter, the setup became easy, the device is in working order and at hand: resistor R9 set the same voltage on the cathodes of the lamps, and the bias resistor set the voltage to 0.55 Volts.However, there is one unpleasant feature of a fixed offset. When the network voltage changes, the lamp mode changes. Once the voltage in the network rose to 250 V and one 6P36S went out of mode, the anode became hot, the quiescent current increased to 80 mA (with the prescribed 55 mA)! Fortunately, I noticed this immediately and turned off the ULF.
I had to install a voltage stabilizer "Ukraine-3", mine produces 215 V. Due to insufficient voltage on the primary (and, accordingly, on the secondary), the voltages on the anodes dropped from the required 330 V to 313 V. I raised the quiescent current to 64 mA ( drop across resistors R12, R13 = 0.64 V).
But time goes by, everyone changes, including me. I got married, settled down, had a child, and lost my free time and money for radio hobbies. And the desire to collect something only intensified. I decided that this amplifier had already outlived its usefulness and I needed to make something decent so that I could put it on a shelf, connect a vinyl turntable with an old jazz record to it and sit pleasantly with cognac and a good book. But there was no time or money for good and beautiful parts, so I collected from what I had.
I took as a basis the circuit of a PP amplifier using 6P36S tubes with a 6N23P driver, because 6P36S cost 10 rubles on our market. The diagram seems simple and understandable, so I assembled it using hanging installation in a few days. I assembled the power supply using a TS-180, took the chokes from a computer power supply at work, and took the outputs from a TAN-104. I picked out filter capacitors from computer power supplies, fortunately there were plenty of burnt ones at work.
I want to warn you right away that I assembled it slowly, there were two main reasons: I had free time only late in the evening (if I had any energy left), an hour before bed, and the lack of parts at hand. After moving to a new apartment, all my supplies remained at my old place of residence, with my parents in a private house.
After assembling the amplifier and checking its operation, the main problem was the housing. I thought for a long time about what to come up with for the case, until one box caught my eye. I remembered that kitchen assemblers brought fittings in this box, so the box was assembled somehow, but, most importantly, the dimensions of the walls (length and height) were the same. I smashed the box, assembled it in a good way and got a normal case, into which both outlets fit while lying down and the power supply could fit into it while standing. I cut out the top cover from the wall of an old system unit (at work I somehow found an old system unit with thick iron). I decided to cover the power supply trans with chokes with the body of a computer power supply unit, having previously cut a hole in it. I cut all the holes in the iron myself, using metal scissors. I drilled crooked holes for the lamps, but under the lamps this crookedness is not visible, and it does not affect the sound :).
I wanted to add different colors to my brainchild, so I took an RGB LED strip I had saved up from a long time ago, picked out LEDs from it (there are three colors in one housing) and put them on glue in the holes in the lamp panels. So that you can change colors depending on your mood, there are three toggle switches on the back, for each color.
The body was cleaned and painted with automotive paint. But it was necessary to immediately coat it with varnish, otherwise chips have now appeared and it needs to be repainted.
I assembled the case, assembled the amplifier, adjusted the anode current and was surprised! I was surprised by the purity of the sound, the presence of bass (this is on TANs) and the lack of background. To be honest, I didn’t expect it, because it was assembled on my knees.
What makes me most happy is the amount I invested in it:
Lamps: 80 rubles (2x6N23P for 10 rubles + 8x6P36S - bought 8 pieces for 60 rubles, chose the best);
- Panels for lamps: 100 rubles for everything;
- Audio input-output panels: 30 rubles:
- Toggle switches for LEDs: 30 rubles;
- Volume knob: 30 rubles;
- Parts (several resistors + interstage capacitors): 100 rubles.
Total for only 370 rubles. The rest is my own, including the paint (I don’t remember why I bought it). Now the amplifier is standing and makes me happy in the evenings with its sound and appearance (especially in the dark). And a vinyl turntable is slowly being prepared for it...
6P45S is a very consumer-grade lamp! For such a healthy bandura, it is made very poorly! Firstly, there is a wide range of parameters. Secondly, the cathode and the terminal are very poorly connected, stuck with some tiny wire that burns out at any overvoltage. I screwed up about 5 lamps. Only the cathode of two of them did not burn out right away, it burned out...in a day! And only one worked for a month. To avoid desoldering, I wanted to parallel two of them, but the filament current of 5 A seemed too high. I used them in my tube high-voltage converter:
http://stalin.flyback.org.ru/tubeflyback.htm
Then I installed a 36 instead of the 45, everything has been working for about a month, the 36 actually calmly holds 600(!) volts and 30 watts at the anode. Reliable as a rock (in a good way).
Arkady Antonov
> anode power 6p36s-20 watts
This may be so, but the lamp calmly holds 27 - 28 watts at the anode... Yes, and thirty won’t scare it
From my experience, 36 sounds more convincing (versus 45)
Pronin
The best-sounding frame tetrodes, in our opinion, are 6P42S lamps WITHOUT A CHAMBER ANODE. They are, however, extremely rare, and the meaning of their production in Svetlana in 1972-1975 is generally unclear. However, they exist, you can find them. Also very good are 6P36S and 6P42s with a white “fluffy” anode. They are also absolutely “indestructible,” apparently due to the coating of the anode.
The sound characteristics of these lamps greatly depend on the modes used.
Therefore, it is difficult to talk about sound outside of these “mode” bindings.
> And anyway, what are the cutoffs for 36 and 42?
I like the 300 V, 125 mA mode for 6P42S and 300 V, 72 mA for 6P36S.
The load is absolutely correct - 2 and 4 com, respectively.
You can wind it on 2 coms, put two sockets and listen to either one 6P42S or two 6P36S in parallel.
There is no need to set 5Ts3S, the sound will be musical, but slow. Install two 6D22S, and achieve musicality by selecting parts.
And in any case, there’s no point in loading the kenotron into space.
Shalin
Well, can anyone tell me the optimal modes for this tube in a triode and the internal resistance and Mu at the operating point... I've been searching - I've listened to the amplifier on these tubes, and they play better with the same circuit design than the EL34 and much better than the EL84. Unlike 6P45S, they stably maintain the mode without floating in thermocurrents at a fixed displacement. Well, in general, do whatever you can to help. As far as I understand, there are no anode graphs for this lamp in nature - I don’t even have “ministry” in my reference book.
Gajdar
Then it’s better to use 6P36S. She plays EVEN BETTER than 6P44S
Shalin
> Hello Alexey. Since you are on the forum, tell me the mode for SE 6P36S and
> given load or a couple of options. Thanks Eduard.
For 6P36S: 330 V, 70 mA, Ra = 5 com.
I like this mode
Shalin
> Just 6P36S is similar to 6P42S of the first releases.
> And 6P36S have less spread and are more stable.
And today I was convinced of this - I was just selecting pairs from 6p36s:
We managed to choose from 15 pieces
1 perfect pair
1 imperfect quartet
4 lamps accelerated
Well, 5 pieces are completely different.
40% yield is a good result
And this ideal couple sounds very beautiful.
Beard
There are Svetlanov and Ulyanovsk 6P36S with light gray anodes - they play better than just gray ones.
Shalin
Better with "42". The 42 anode is slightly larger than the 36th and similar to it, and the 45th is
1.5 times higher and at the ends on both sides there are three rectangular holes of approximately 6x6 mm. It seems that A. Shalin posted photos of the correct 42 somewhere. 36th so - if the anode is almost white and “fluffy” or light-light gray and again “fluffy” - then you need to save it, even a used one.
HRYUN
You need to turn it on for 2-3 minutes, just long enough for the lamp to warm up. They feel normal. This is not yet serious overheating. Forty-five is a strong lady. They only collect them poorly.
The meaning of such a check is as follows. In a beam tetrode, both grids have the same pitch, and must be installed so that the grid threads are exactly opposite each other. This is how rays are formed. 6P45S and 6P36S have almost the same cathode-grid design - 4 frames with welded wires. Two frames on one side of the cathode, two on the other. It turns out two sets of rays in different directions from the cathode, to the opposite halves of the anode. Let's assume that on one side the frames are not aligned exactly. Then the beam current on this side will be less than on the other, and the anode half on this side will heat up less than the opposite. Half the lamp is a beam tetrode, and half is a conventional one. And these two tetrodes, different in characteristics, are connected in parallel. In principle, you can use a pyrometer to measure the temperature of different parts of the anode, but in the absence of one, heat the lamp for a short time until redness appears. If it turns red more on one side, it is a clear sign of a manufacturing defect. It will work fine in horizontal scanning, but it’s better not to use it in sound. There are approximately 80-90% of such defective lamps.
Oleg
Message from KHRUN
And the truth is HORROR...! I’ll tell you my terrible secret: somehow (long ago)
my 6P36S (vintage) worked for quite a long time in 250 V mode,
160 mA (40 watts, however...) with auto-bias (more precisely, almost with auto-fix, but then no one knew that it was auto-fix). And nothing, they remained alive.... Gridlik was, it seems, 51 kOhm.
Likewise. Only my used Svetlanov 6P36S worked for several months in the 100 mA 400V mode, and with a fix.
Dalka
And this is precisely the reason for the bad sound... I came across 45s that turned red on one side at 40 watts at the anode, maybe a little more. They immediately left the trash. How much power can the selected lamp dissipate before turning red?
Sergey Z
Those selected at 60 practically do not turn red. Only in complete darkness is a slight glow noticeable. Indeed, this is a good alternative to 6S33S. It is much easier to swing and much more linear.
Oleg
I looked at the triode I-V curves of the 45th.
I found one mode:
250V, 180mA, -50V on grid.
Ri = 290 ohms, Ra = 2380 ohms, alpha = 8.2.
Uam = 181 v, Iam = 76 mA,
P~ = 6.88 W.
Linearity in this mode is very high.
In the 250 V, 240 mA mode, it will not be possible to make a load of more than 1242 ohms, since then the right half of the load line will even go beyond the power curve of 60 (!!!) watts.
In a word, whatever one may say, the use of 6P45S with an anode power of more than 45 W is doubtful for me...
Shalin
6P36S are very vibration-resistant lamps with strong frame grids, they have a small microphone.
Shalin
IMHO, in comparison with the 6P36s, the 6P44S is first captivated by the novelty of the sound, then, after listening,
you understand that the sound is “prickly and rough”, but subjectively there are more highs,
when measuring a longer tail of harmonics, comparative measurements were carried out only
one output stage on different lamps, all other things being equal.
As for the spread, Svetlanov 6P44S have closer parameters than 6P36S,
for 6P44S the average spread is up to 30-35%, 6P36S up to 50%.
Everything is learned by comparison, but previously unused, selected 6P31S, IMHO have the most
natural sound, close to the sound of 2A3 in the midrange and treble.
Manakov
Dmitry, one of my 6P36S amplifiers has been working for 8 years at 20 watts on the anodes without replacing lamps. My friend has three years at 27 volts.
Somewhere Hryun pointed out that in the forced (36 watts at the anode) mode the Ri of the 6P36S drops to 450 ohms.
I myself try not to “torment” the 6P36S with more than 28 watts.
Shalin
Dmitry, this means that the lamps were not very good
Good 6P36S behave perfectly with 33-100 ohms in the grid. But anti-excitement measures are definitely needed, that’s true.
I did not test the 6P36S with more than 32-33 watts of anode power, but my friend in the amplifier with a 6P36S (my build) tinned it at 37 watts, and lives normally with a fixative and even without an anode stub.
Shalin
Push-pull tube stereo amplifier based on 6P36S tube NEW! February 25th, 2011
Finally got around to the article. Let's begin.
Tube sound differs from semiconductor sound for the better. There are many of them, but I won’t tell you. The only negative is that making a lamp is a tedious and difficult task. But it's worth it. To avoid any questions, you should read it.
I have already assembled the acoustics, so that there are no disputes, I will say right away that it has a sensitivity of 102 dB - just right for a lamp!
http://community.livejournal.com/ru_audiomania/1540.html
Let's move on to us. It is assembled according to a push-pull (PP) circuit using 6n23p lamps in the first stage and 6p36s in the output stage.
Scheme
Scheme by Sergei Sergeev with my modifications.
The 23x filaments will be connected to ground through 150 ohm connections from each pin. Also chokes for the anode. Well, I’ll put more electrolytes.
The setting consists of setting the voltage on resistors R12 and R13 to 0.55 volts.
Transformers
The first thing I did was make them.
For the network and two outputs I used iron from TSSh-170 transceivers with a plate thickness of 0.35 mm. I rebuilt the frames, although it was not easy.
Output parameters:
We divide the frame with the middle cheek. We swing the halves in different directions.
On each half:
Primary - two sections of 560 turns (10 layers of 56 turns) of PEV-2 0.355 mm wire.
R act of the primary - 98 ohms.
The secondary - between them - 112 turns of the same wire in two layers, taps from the 56th and 79th turns for 4 and 8 ohms, respectively. 112 turns - for 16 ohms.
There are three such secondary units in parallel on each half.
R act of the secondary - 0.88 ohm. Given - 352 ohms.
We connect the primary windings crosswise in series, the secondary windings in parallel. For more details, see G. Tsykin’s monograph (it’s cunning, by the way).
In total, the frame has 2240 turns in the primary winding and 112 in the secondary.
The iron naturally fits across the roof without a gap.
Each trance lasted 12 hours. Chewy. But what is the result:
I just exhausted the network to get 280V ~.
We end up with 360V at the anode, taking into account the drawdowns.
For the 23rd we wind two separate filaments.
To heat the exhaust we need 8 amperes. A standard winding will provide them without any problems. Near trans offset:
Chassis
I thought out the location and design myself. The chassis and all the trans boxes will be stamped and welded automatically at the factory, otherwise it will simply be ug (besides, I have long wanted a normal body).
(in the photo there is a model with another trans and a couple of chokes)
For reliable ground and glow I use these beautiful tires:
Well, for fastening all kinds of dowels I bought:
A trial prototype run gave the go-ahead, I liked everything, the sound was excellent!
The video is a MUST watch!!! You can see everything there.
I specifically asked a friend for a camera!
It’s a pity YouTube has degraded the sound quality, so it’s advisable to watch at least 720p, and better yet 1080p!
While there is nothing to write about, the chassis is still being welded at the factory. When it appears, I will immediately continue the article! In the meantime, write your thoughts, impressions, opinions here. I will be happy to answer your questions.
======================================== =====
Continuation.
The chassis has arrived! Everything is as I said, 2mm steel, holes only for sockets:
Then the security officer arrived in time. From the c1-1 tube oscilloscope, which is great for my purposes:
The panels will be flush:
The main idea is without a single visible bolt.
Therefore, the method of soldering flat head bolts to the chassis was undertaken. It is necessary to clean the surface very well so that it does not fall off and holds tightly. Then solder with a 100-watt soldering iron.
As a result, all elements of the circuit are located on special panels for wall-mounted installation.
The anode electrolytes are located on special holders that are thermally insulated from the chassis.
The setup involves setting the lamp and balancing modes.
Then it was decided to add shunts to the anode power supply at KBG 500V 5uF.
Sound.
The sound is as expected at a very high level. Low THD 0.5% at 28W. This will be enough for my speaker, given its very high sensitivity.
I'm currently planning my weekend caps.
