THIS PROJECT HAS NOT YET BEEN COMPLETED AND/OR TESTED.

THE INFORMATION YOU WILL FIND HERE COULD BE INCOMPLETE OR INACCURATE. THE WEBMASTER, PROJECT COORDINATOR, PROJECT CONTRIBUTORS AND WEB SPACE PROVIDER EXPRESSLY DISCLAIM ALL LIABILITY FOR INJURY OR PROPERTY DAMAGE RESULTING FROM THIS INFORMATION.

High voltages will not just shock you, they will kill you.
Use extreme caution while building, working on and testing.
Always THINK SAFETY! Only YOU can prevent accidents.

If you are in any way unfamiliar with high voltage circuits or are uncomfortable working around high voltages, PLEASE DO NOT RISK YOUR LIFE BY BUILDING THEM. Seek help from a competent technician before building any unfamiliar electronics circuit.

USE AT YOUR OWN RISK: THE WEBMASTER, PROJECT COORDINATOR, PROJECT CONTRIBUTORS AND WEB SPACE PROVIDER EXPRESSLY DISCLAIM ALL LIABILITY FOR INJURY OR PROPERTY DAMAGE RESULTING FROM THIS INFORMATION! ALL INFORMATION IS PROVIDED 'AS-IS' AND WITHOUT WARRANTY OF ANY KIND.

Wear rubber soled shoes to help keep any accidental shocks from passing through your body. Make sure the unit is unplugged and the capacitors are completely discharged before your body or tools come in contact with the equipment you are working on. (Electrocution - A widely misused term, it refers to death by electrical shock. People often say that they have been electrocuted; that is just not possible, they would be dead. You can only be electrocuted once....)

Capacitors can vent (commonly know as exploding). Wear safety glasses when working with electronics, especially when working with capacitors. Wear two pairs the first time voltage is applied to a capacitor after soldering. If you care about what your face looks like, wear a complete face shield. If possible, put a nonflammable plastic box over your work the first time you apply power.

Wear long cotton pants. Solder drips burn through polyester and hurt for weeks. Don't wear pants that cost too much to be ruined with a burn mark.

If possible, work with a friend who knows CPR and knows how to pull the plug from the wall if you forgot to do that first. Do not work on a system, under any circumstances, while it is plugged in. Don't just turn it off and think that's enough. Others have done this and fried things (or themselves!) by accidentally turning the system on while working. You could end up in the hospital or in a wooden box with just the slip of a screwdriver.

Work with a full fire extinguisher in view. Do not put the fires extinguisher where you would have to reach across equipment to get to it.

The case of a capacitor can have voltage on it. Check it with a working voltmeter on both AC and DC before attempting to touch it.

Make sure you can pull the plug from the wall without having to reach over either the equipment you are testing or the test equipment.

Make sure you can exit the room rapidly without passing your equipment or tripping over piles of stuff on the floor should you need to run for it.

Make your own safety assessment. Your design and craftsmanship must be safe for both you and those who enter your house. Do you have any friends, relatives or acquaintances who insist on touching things when they say "what does this do?"

Both old and new parts can be defective. Test them before installing them.

Make your own risk assessment. Read the application notes for the parts you are using.

Watch out for tantalum capacitors. When they vent, they can spew molten tantalum!

• Keep one hand behind your back. A very dangerous path for high voltage is from hand to hand, with your heart in between. It doesn't take all that much to stop your heart at these voltages.
• Use proper probes. About the only thing you should be doing to a live circuit is checking voltages or signals. Get clamping tips for your probes. Clamp the ground clip to ground. With the hand that is not behind your back, use the proper insulating probe tip on your voltmeter or scope. Be sure your voltmeter or scope is rated for at least 500 volts for this project.
• Disconnect the power and DISCHARGE THE CAPACITORS before doing anything to the wiring. The capacitors in this project can give you a high-voltage jolt at much more than 27mA, and can hold that charge for a long time (e.g., an hour or more). Do NOT discharge the capacitors with the crude method of shorting the positive side of the caps to ground with an insulated screwdriver.A better method is to bend a 10kΩ/2W resistor into a "U" shape, tape the body of the resistor to the end of a stick, and use that to discharge the capacitors. Measure the capacitor voltage with a meter to be sure it has drained all the way down: they don't discharge all at once, it takes 15 - 20 seconds, or more.

  And don't forget to discharge the audio output capacitors, too!

   

If you use the information in this site to kill yourself, your friends, family members, acquaintances, total strangers, pets, electronic devices or burn down your house, it is not my problem.

I will nominate you for a Darwin Award.

Proceed at your own risk.

Everything started surfing the Internet looking for a cheap LCD display... and stepping on an E-bay auction for a set of Nixies. (I actually bought them... and built a peculiar Nixie clock; but that is another project, detailed elsewhere on this site).

From then on I became more and more interested to the old vacuum/gas filled tubes technology...

After finishing the Nixie Clock I started thinking to make something more difficult: an audio amplifier. BUT... I wanted for the first project to avoid the most expensive and difficult to find parts: the transformers, especially the output ones [ I did not know yet that there is a cheaper (non-otl) way... more on this later ].

By then I was trying to understand this new (to me) 'old' technology, and I started studying a lot of documents found on the Internet (see the References). In this search of knowledge, I found this wonderful site: http://headwize.com/, a real mine of ideas, with a pool of competent and helpful enthusiasts.

This is a project found on HeadWize, heavily modified in some parts; in my opinion mine it is easier to realize (having the components I used by hand), but basically it is the same. You could find the original one HERE.

[Well... 'choice' is not the right word here for the main components (the tubes)! I bought on Ebay a set of 8 6DJ8 tubes, and received 8 6N1P-EV. Looking at the feedbacks of the seller I noticed a lot of negative feedbacks for the same reason, and people that shipped the items back to be refunded did see neither the items nor the refund... so I kept the 6N1P-EVs. Be extra cautious when buying on Ebay, and always look at the seller feedbacks before bidding! I discovered later that 6N1P-EV are "better", somehow, than 6DJ8... and tried to make good use of them... It could be useful to look at the original 6N1P-EV datasheet, that proved to be somewhat hard to find... and for empiric measurements you can look at this page ]

Due to the 'choice' above mentioned, I needed a high filament current draw than expected (600±35 mA * 3 ≈ 1800 mAmin @ 6.3V), and A LOT of B+ voltage (the original project required 350V, but I simulated a 300V B+ and the results were close enough to perfection to satisfy me; the nominal Ua for 6N1P-EV is 250V). 6N1P-EV is a rugged military-grade tube with very low microphonics (admitted vibration load 6g!) and extended service life (≥ 5000h). This is the pinout, seen from the bottom:

Due to the prerequisite not to have any tube-dedicated transformer in the project, and knowing that the B+ current must be very low (around 30 mA), I thought to (carefully) use an oversized step-down transformer in reverse, coupled to a conventional switching power supply design.

Throughout the project only high quality components must be used; do not be fooled by 'audio' components though... it is enough in my opinion to use precision resistors, low ESR electrolytic capacitors (like the ones used in switching power supplies, for example), and polypropilene for non-polarized ones. Better to be extra cautious in using exactly the same components (model and batch) for both channels.

Where an high precision resistor is needed, you can reliably obtain it from lower precision ones... errors tend to statistically cancel each other, so if you need a 43kΩ,1% you could use:

130kΩ 5% // 330kΩ 5% // 330kΩ 5% // 330kΩ 5% // 240kΩ 5% = 42.996kΩ ≈1%

(And you gain an higher wattage, too. Probably you will need to double check the obtained value with a precision ohmmeter... and change some resistor until the measure is correct).

There is a drawback... if one of those resistors fails, the total value will vary unpredictably. but I cannot see a reason for a quality resistor to fail if correctly dimensioned.

I decided to use low impedance cable, with silicone insulation, to connect the tubes. I don't know if this is really affecting the musical performance, but who knows? This is something you have to decide before wiring everything... and silicon cable does not suffer from heat and age...

The main question here is: how to obtain a 30mA 300V B+ without a tube-dedicated transformer, PLUS a 2A 6.3V heater supply, with the minimum complexity and cost?

There are virtually infinite solutions... and i have chosen this one:

The underlying principle is to make the 555 oscillate, thus generating an alternating current, and feeding with it a transformer wired in reverse; I extracted one from a big wall-wart (primary: 230V 50Hz, secondary: 13.5V 1A), obtaining exactly 300V at U3.

R3 is not there for filtering purposes... it is a bleeder... due to its high value it does not really affect the charging of the capacitor C5, but guarantees that after a reasonable amount of time the capacitor will discharge if the circuit is powered off. I did not like the idea of opening the box after a week for some checks, forgetting to have a 385V capacitor still fully charged inside!

Be extra careful in choosing the transformer; it must be oversized not to suffer from the stress being used backwards; but do not overkill, we need "only" 30mAmax!

As specified in the original HeadWize project, if you want a dummy load to use when checking the power supply, you will need to cobble together a 13k ohm load capable of dispersing 10 watts.

Notice that U4 is NOT connected to ground... this means that heaters and B+ will not have the same ground plane. But the heaters share the same ground with the 7815...

Everything is powered by a (recovered from trash and repaired!) laptop switching power supply, 20V 3.5A.

To reduce the current requirements and the wiring complexity, I choose to wire the heaters IN SERIES. The same 20V that powers the B+ supply will be wired to the heaters... but with 3 tubes, if we connect the heaters in series, we will achieve 6.6V, not 6.3V. So I connected a 1Ω, 20W resistor in series to the heaters...

The charging of the 220uF 300V capacitor is essential to the stabilisation of the B+ supply, but it does not happen instantly! Moreover,  the cathodes need to be fully warmed by the heaters before applying B+ to the plates... if B+ is applied to the plates before the cathode is warmed up, internal arcing could occur, damaging or destroying the tubes.

There are different approaches to this problem; being used to programming, I decided to use a small Microchip PIC processor (12F508A) to implement the following simple logic:

1) Wait for the toggle of a button, with a simple debounce algorithm

2) If the device is currently powered off,

a) Actuate a relay that powers up the heaters; switch on a green led, and pulse a red led

b) Wait 1 minute

c) Actuate a relay that powers up the plates; switch on the red led, and restart from 1

3) If the device is currently powered on,

a) Power down the plates; pulse the red led

b) Wait 2 seconds; switch off the red led

c) Power off the heaters; switch off the green led

d) Wait 1 minute, and restart from 1

Notice that if the amplifier is suddenly unplugged, the relays will open, thus allowing a safe restart when plugged in again...

This is the schematic:

The RLY2 will switch on also the B+ power supply... after 60 seconds the B+ capacitor will be charged, and the cathodes will be fully heated.

Here is the (very simple) PIC code, written in CCS C:

#include <12C509A.h>
#use delay(clock=4000000)
#fuses INTRC,NOWDT,NOMCLR

#define GRNLED PIN_B0
#define POWER PIN_B1
#define REDLED PIN_B2
#define BTN PIN_B5
#define BPLUS PIN_B4

#ZERO_RAM

void delay_seconds(int n,int nLED)
{
    short bLed = 1;
    for (;n!=0; n--)
    {
        delay_ms( 1000 );
        if (bLed == 1)
        {
            nLED == GRNLED ? output_low(GRNLED) : output_low(REDLED);
            bLed = 0;
        }
        else
        {
            nLED == GRNLED ? output_high(GRNLED) : output_high(REDLED);
            bLed = 1;
        }
    }

    nLED == GRNLED ? output_high(GRNLED) : output_high(REDLED);
}


void main()
{
static int16 tmrAction = 0;
short bDeviceOn = 0;
short bGoingOn = 0;
short bGoingOff = 0;

    setup_counters(RTCC_INTERNAL,WDT_18MS);

    output_low(BPLUS);
    output_low(POWER);
    output_high(REDLED);
    output_high(GRNLED);

    delay_ms(1000);

    output_low(REDLED);
    output_low(GRNLED);

    while(1)
    {
        if (bGoingOn == 1)
        {
            // SWITCHING ON...

            // POWER UP
            output_high(POWER);
            output_high(GRNLED);

            // WAIT 1 MINUTE FOR THE HEATERS AND B+ CAPACITOR
            delay_seconds (60,REDLED);

            // B+ ON
            output_high(BPLUS);

            bDeviceOn = 1;
            bGoingOn = 0;
            bGoingOff = 0;
        }
        else if (bGoingOff == 1)
        {
            // POWER DOWN...

            // B+ OFF
            output_low(BPLUS);

            delay_seconds (5,REDLED);

            // POWER OFF
            output_low(POWER);
            output_low(REDLED);

            // WAIT 1 MINUTE
            delay_seconds (60,GRNLED);

            output_low(GRNLED);

            bDeviceOn = 0;
            bGoingOff = 0;
            bGoingOn = 0;
        }
        else
        {
            // CHECK OUT THE BUTTON
            if (bGoingOn == 0 && bGoingOff == 0)
            {
                if (input(BTN))
                {
                    tmrAction++;
                }
                else
                {
                    tmrAction = 0;
                }
            }

            if (tmrAction == 500)
            {
                if (bDeviceOn == 0)
                {
                    bGoingOn = 1;
                }
                else
                {
                    bGoingOff = 1;
                }

                tmrAction = 0;
            }
        }
    }
}
 

To understand how to dimension the components around a tube we must talk a bit about characteristic curves.

First let's choose the B+, trying to keep it low, possibly in the center of the admitted plate voltage of our tube. We choose 300V.

One tube parameter that can be calculated from values on the Ep - Ip (Plate Voltage - Plate Current) curve is known as plate resistance, abbreviated as Rp. In a properly designed electron tube, there is no physical resistor between cathode and plate; that is, the electrons do not pass through a resistor in arriving at the plate. You may have wondered, however, why the variable dc voltage source of a tube does not blow a fuse. Doesn’t the plate circuit appear to be a short circuit-a circuit without a load to limit the current? The fact is, there is a very real, effective resistance between cathode and plate. It is not lumped in a resistor, but the circuit may be analyzed as if it is. The plate resistance of a given tube, Rp, can be calculated by applying Ohm’s law to the values of Ep and Ip: Rp = Ep / Ip

But, thanks to Ohm's law, if we want Voltage, and we have Current, we can add Resistance (E = I x R). In other words, run the plate current variation (caused by the voltage on the grid) through a resistor, and cause a varying voltage drop across the resistor. VOLTAGE GAIN is calculated by dividing the output signal voltage by the input signal voltage.

Typically for a triode, we want to load it many times its plate resistance. 6N1P has a nominal plate resistance of 4400 Ω so if we load it 5x, this means a plate resistor of 22 kΩ.

Now let's look at our typical plate characteristics:

2) Plate Voltage = 0. This means MAX CURRENT flowing through the tube. To have max current flowing through, the only way is the plate voltage is at zero volt. It's like saying the tube has shorted itself to ground, so we have a short circuit where between B+ and GND, the only resistor we have is the plate resistor of 22 kΩ. So we now have 300V = Current * 22 kΩ. Current should therefore be 300 V/22 kΩ = 13.63mA. (Note that we could proceed iteratively to find the plate resistor value that better fits the plate characteristics of our tube... obeying the tube maximum electrical ratings)

We can also proceed to calculate how much distortion is there....

In circuits using cathode bias, the cathode is made to go positive relative to the grid. The effect of this is the same as making the grid negative relative to the cathode. Because the biasing resistor is in the cathode leg of the circuit, the method is called CATHODE BIASING.

Navy Electricity and Electronics training course - Electronic Emission, Tubes, And Power Supplies.pdf     (1,5 Mb)  (Approved for public release; distribution is unlimited)

A very comprehensive text I have is:

Vacuum Tubes - Spangerberg, 1948.pdf     (73 Mb!!!)    (I am verifying if it is still covered by copyrights... I hope to put it online)

And then there is THE BIBLE:

"Radiotron Designer's Handbook", edited by F. Langford-Smith, Fourth Edition, 1953, 1422 pages. I have it, but I cannot put it online due to copyright issues. Be forewarned that if you want to read it you will have to deal with heavy mathematics... but if you succeed in going through it, you will really know everything about tubes.

I do not like AT ALL the messy wiring normally used in tube amplifiers, or the old rule "place components as random as you can to avoid interferences"; one thing is to appreciate old technology, another thing is to use it exactly as if we were in the fifties... so I assembled everything on a perfboard

(excluding tubes, but I am working on it, and soon a double-sided PCB will be engineered), following some simple rules:

1) Do not place components too close together, to avoid mutual interferences

2) Place high voltage sections far from low voltage or sensitive ones

3) Phisically separate the power supply from the amplifier

4) Shield anything you can reasonably shield

5) Use shielded, high quality cables for input connections and anything going outside the chassis. (I usually use very good cables from scrap equipment, to save money...)

6) Do not be scared by a tight connector going to the tubes, giving a proper insulation of every connection

7) Think about easy maintenance and repairing; write any useful information (by hand or by any other means) on (or near) the board where components are placed. After a couple of years you will remember little (if anything) of the details of this project... and consider that someone else could have to take care of your work (see the disclaimer and safety information... ok, I am joking...). In any case, write down everything in a lab diary, or on a web page, if you prefer... ;o)

8) ALWAYS CONNECT THE CHASSIS TO GROUND, USE FUSES, AND THINK ABOUT HOW COMPONENTS MAY FAIL IN A DANGEROUS WAY

9) Protect the tubes! (think about your amplifier falling and shattering the tubes, and someone rushing to it and coming in touch to the plates still connected to B+...)

Of course these are a personal opinions; I do not want to arise any flaming around... maybe I really am a bit too 'modern' in my views...

This amplifier has has a grounded cathode input stage, and a push-pull White follower output stage, with variable feedback to optimize driving low-impedance headphones. Here only one channel is represented; three tubes will be needed to have two channels (remember that 6N1P is a double triode: V2 and V3 are in a single tube... V1 is half 6N1P!)

The simulation has been conducted with CircuitMaker 2000 , with this SPICE model of 6N1P tube:

 

*Vacuum Tube Triode (Audio freq.) pkg:VT-9 (A:1,2,3)(B:6,7,8)
.SUBCKT X6N1P 1 3 4
B1 2 4 I=((URAMP((V(2,4)/34)+V(3,4)))^2.2)/1400
C1 3 4 3.2E-12
C2 3 1 1.6E-12
C3 1 4 1.5E-12
R1 3 5 10E+3
D1 1 2 DX
D2 4 2 DX2
D3 5 4 DX
.MODEL DX D(IS=1.0E-12 RS=1.0)
.MODEL DX2 D(IS=1.0E-9 RS=1.0)
.ENDS X6N1P

 

 

Point A is the input signal, varying from -300mV to 300mV, 1Khz

Point B is the output signal; here it is attenuated, but it can swing from -4V to 4V

Point C is the input to the White follower stage from the grounded cathode stage

Point D is the output from the White follower stage

Point E is the followed signal, input to the second triode of the follower stage

HOME