12VDC EF97 shortwave superhet

A while ago, someone asked me to make a receiver with valves that could be copied by kids with the objective to generate interest in Radio and history. It was important that the set does not carry high voltages and there was little or no alignment to be done, no measuring equipment available. So I started to experiment with the EF97, a valve that is very happy with 12 VDC at the anode. I did like to use 2 valves only due to availability and to connect the filaments in series for 12 VDC This article is the result, the radio has been reproduced quite a few times by some kids and radio hams. In order to have good selectivity without (for many kids) difficult to control regeneration, I decided to make a superhet. But IF transformers could not be found anywhere for a low price and always the same, so I made the transformers my self from simple small 4700 uH chokes. That proved to worke very fine and there was no alignment required, an other advantage. The 6 VDC required for the second grids of the valves could be found at the interconnection between the filaments. Less wiring and resistors. I did set the radio at the shortwave band from 15 - 22 MHz. But by inserting ferrite cores in the antenna coil and the oscillator coil, the band could be lowered to 6 MHz, all without changing or switching coils or caps. If one likes, it is off course possible to set the receiver at MW instead of Shortwave using other coils and tuning caps. But for the kids shortwave was most interesting, one can receive easily very far stations as China, Africa, Indonesia, and Rusia. Schematic diagram. The receiver starts with a tuning circuit with a small (40 pF) tuning capacitor, directly connected to the grid of the first EF97 In order to prevent high loading of the circuit by the antenna, the coupling coil to the antenna has few turns. I use a ratio of 1:15. That keeps the Q high in order to reduce mirror sensitivity. The first EF97 is also the mixer. The LO signal is injected at the third grid. That gives an excelent isolation between LO and input to prevent LO pulling when the input circuit is tuned. The IF transformers use two small 4700 uH chokes with a centre to centre distance of approx 8 mm. That gives a nearly optimum magnetic coupling. The tuning capacitors are 120 pF. The transfromers are mounted using a small strips of PCboard without copper. The construction can be seen at the pictures. It is important that all 4 chokes of the two transformers are of the same type and model. The second EF97 is the IF amplifier. The IF transfromers have a high impedance, so the loading should be minimised. Therefore, the load after the Schottky detector is 2.2 MOhm. In order to keep the distorsion of the detector low, the AVC for the IF amp is coupled via 22 MOhm. Just for sake of simplicity and to prevent more hard to find valves, I used JFET's for the LO and the LF amp. But if one wishes to keep the receiver pure valve, these can be EF97 as well. The AF stage is merely a impedance transformer. About any headset can be connected. If more volume is required for low impedance headsets, the source resistor of 3k9 can be lowered. Even a loudspeker with an impedance of 50 -500 Ohms will give satisfacory results when the source resistor is replaced by a 1 kOhm potmeter to allow volume control. About the LO nothing special there. It is highly advised not to tune a wide frequency range, otherwise one needs bandspreading or a mechanical reduction drive to tune. The selectivity is quite high. The sensitivity is quite good, a 1uV AM signal is audible. Enjoy experimenting Regards Nico ps I will sent the schematic diagram separately, don't want to loose my text due to very bad internet overhere Attachments:

• 12VDC EF97 based superhet (31 KB)
• back chassis (36 KB)
• underside (68 KB)
• IF transformer coils (81 KB)
• IF transformer caps (56 KB)

To thank the Author because you find the post helpful or well done.

Dear Nicolaas

Interesting solid state/vacuum tube hybrid design. After reading the description and looking at the circuit diagram, i have some questions:

With the 4700µH inductors (RF chokes) and 120pF capacitors, the intermediate frequency is 212kHz, which is unusually low. My guess is that since RF chokes typically have a relatively low Q-factor, the IF needs to be quite low for the (absolute) bandwidth of the IF filters to be approximately the required 9kHz. Is this assumption correct? What is the Q-factor of the RF chokes and the bandwidth of the IF filters? Can you post the frequency response curves?

With an IF of only 212kHz, image (mirror) frequency rejection becomes quite difficult and requires a low bandwidth (high Q-factor) of the input tank. You mentioned a "high" Q-factor of the input tank due to the loose antenna coupling. What is the Q-factor of the input tank at 6, 15, 22MHz with a typical antenna connected?  
 
Back to the IF filters: The tolerance of good RF chokes and capacitors is usually +-5%, giving rise to a maximum deviation of the resonant frequency of each single tuned circuit of approximately +-10kHz. This is even more than the (optimal) bandwidth of the IF filters of 9kHz. Did you hand-pick matching chokes and capacitors to build the IF filters or did the components in use just happen to have (almost) identical values?

Regards,

Jochen

 

To thank the Author because you find the post helpful or well done.

The Russian Rod Pentodes can perform with NiMH AA LT and 3 x 6LR61 HT very well. 1j24b is 11mA filament

The ECH81 = ECH83 and as ECH81 is common. It works from 14V HT.  Grid is then 0V to +1V, you can add 7V 300mA bulb / lamp and also run the filament from same 12 V to 14V supply. Or add a 2nd one. It need not be used as a frequency changer. But can be Pentode RF amp and triode AF amp. Then a ceramic earpiece across the triode anode load. Add 1N60 detector (used since 1946).

If using two then you can just have more gain or a simple superhet. The first as Mixer/OSc, the 2nd Pentode as IF amp and triode as RF Preamp. The ferrite rod and tuning cap from a €6 pocket radio can be used.

I have used the MW coils of cheap radio (with core cut to coil length and 470pF cap as IF coil. Two about 8mm apart make a good IFT. tune by moving core and lock with wax.  You need two sets for superhet.

At 14V you can only drive a crystal earpiece. At 48V HT though you can use a cheap miniature mains to 9V transformer as audio out to drive an 8 Ohm speaker. Especially if using a 1j24b rod pentode.

 

To thank the Author because you find the post helpful or well done.

@Jochen Bauer:

"Back to the IF filters: The tolerance of good RF chokes and capacitors is usually +-5%, giving rise to a maximum deviation of the resonant frequency of each single tuned circuit of approximately +-10kHz. This is even more than the (optimal) bandwidth of the IF filters of 9kHz. Did you hand-pick matching chokes and capacitors to build the IF filters or did the components in use just happen to have (almost) identical values?"

 

To overcome this problem, use trimmer capacitors for C9, C10, C15 and C16 to properly align the IF strip. If I were you, I should not bet on the values of the chokes and fixed capacitors of the IF filters.

I see that the screen grid to which the signal from the Local Oscillator is fed, has been shunted to ground for RF by C8 with a value of 10 nF. I would omit this capacitor, otherwise there will almost be no LO signal on the screen grid and the mixer performance will be very poor. And it is very sensible to use a buffer stage bteween the Local Oscillator and the mixer stage, otherwise the Local Oscillator will be influenced by the RF signal which is fed to the control gird of the mixer pentode.

To thank the Author because you find the post helpful or well done.

Dear Jochen

I did test some transformers when I did design the receiver, already quite a few years ago. I found that it was not required to handpick or select the coils, I just bought 50 and the coils. All where within a few kHz the same. The 5% tolerance mainly appears when you have different batches. 

Also, the objective was that kids could reproduce the receiver. Off course they do not have equipment nor the knowledge to align the transformers. 

I used 2.5% caps though, in order to keep the frequency tolerance within PRACTICAL limits. 

I think that the bandwidth of the IF is approx 8 - 10 kHz when coupled KQ=1 The curve shows a nice critical coupled filter, sometimes with a slightly sloped top when there is a tolerance problem. So the Q will be in the order of 20 - 30.

The Q of the inpurt circuit with a few meters wire as antenna (sufficient to receive very many stations) is in the order of 100 at 10 MHz. So no great mirror rejection, but very noticable. When I tune the input, you can clearly distinguish two noise peaks. And when there are two stations, one at the input and one at the mirror, it is very well possible to select one or the other without too much interference by tuning the input only. (when one of the stations is not extremely strong)

The sensitivity of 1 uV is high, with the disadvantage that a larger antenna will give IM and more interference of mirrors. 

I think that the objective, a simple radio with valves that can be reproduced easily by kids has been reached. And normally they do not have big antennas anyway. And it has been reproduced quite a few times with good results.

Kind regards and thanks for the interest. 

To thank the Author because you find the post helpful or well done.

Hello Michael

Thenks for the info. I did not know that the ECH 8X worked with such low voltages, interesting.

The 12 VDC is chosen because these powersupplies are very low cost (27 MHz) and many small lead acid batteries are available. So the set should funcion from off 10 VDC.

Also, nowedays it is really difficult to get crystal earpieces. So the design is made to accept almost any kind of headset, also the ones used for stereo players. Remember, it is for kids wich do not have the knowledge to find the correct parts. So the difficult parts I bought in quantity, the IF coils and caps, the EF97, the tuning caps and coil formers. I still have approx. 20x EF97 laying around . 

I have a few of the russian rod tubes, you made me curious to try these one day!! But for projects like this, the availability is not sufficient and the kids may easily toast filaments...And soldering a 7 pin socket is much easyer for them.

The EF97 performs very well as mixer when the LO is injected at the third grid. The receiver has an overall gain (including the FET impedance transformer, off approx 120 dB, so quite impressive for only two tubes and a low gain FET.

Kind regards

Nico

 

To thank the Author because you find the post helpful or well done.

Dear Daniel

Thanks for the comments, appreciated.

But the LO is fed to the third grid, not to the screen grid. This especially to have a very high isolation between the LO and the input  circuit. And it is indeed very high, in practice no pulling can be experienced at all.

The EF97 performs quite well as a mixer this way and the overall gain is high. The sensitivity is 1 uV for a audible and understandable AM signal. 

The objective was the make a design that could be reproduced by kids. So tuning the IF's with trimmers will make a better IF curve, but kids do not have the tools nor the knowledge to align an IF. That was the reason to use fixed coils with a low IF frequency and low Q in order to get a practical IF system without tuning.

Kind regards

Nico

 

To thank the Author because you find the post helpful or well done.

NOS Russian rod tubes are very common*. They are under €2 in quantity inc. post from Russian, Ukraine and Bulgaria on eBay and 1977 to 1990 date.

I've never even seen an EF97 in 40+ years of using valves (tubes).

The ECH83 (totally = ECH81) was used in virtually every car radio that used Germanium output transistors and car 12V for HT (no inverter or vibrator).. They are common and better than EF97 for a mixer/Osc. No valve can drive a speaker at 12V HT.

If using ANY semiconductors I'd use a LM386 to drive speaker or regular earphones. But at 32V to 48V and a small 9V mains transformer as audio transformer you can drive 32 Ohm Stereo phones in series (64 Ohms) with the triode of the ECH81. With a Rod Pentode you can drive an 8 Ohm or 16 Ohm speaker with 48V HT and 1j29b.

With a trickle charged NiMH C cell or AA cell  (like principle of DEAC in old German sets) there is not much risk of damage to Rod Pentode filaments. I've used cheap miniature plastic screen terminal strips to make a Rod Pentode radio with no soldering.  Using I/2 an IC socket strip works too.

I've replaced DF96 DAF96, , DK91, DK92, DF97, DAF91 and DK96 with 1j24b. Two make a Mixer/Osc, with one g2 to anode as triodised osc, g3 = f-, this then feeds direct (no capacitor) to g2 of the RF /IF mixer.

DL94, 1j29b and DL96 replaced by 1j29b with only 1/.2 filament connected. All parallel filament only.

Edit:

* Though production ceased in 1991, one estimate is that there are millions of Russian Rod Pentodes available. Certainly they are cheaply available at 1000 off and it's worth buying 10 of each needed. 1j24b is 11mA filament. Push Pull can do 500mW, or 150mW single ended. The 1j29b is much more gain and maybe 800mW single ended or  1.5W class B push pull. It's about 26mA @ 2.4V or 53mA @ 1.2V Operation over 100MHz is possible.

To thank the Author because you find the post helpful or well done.

Dear all.

I found an error in the values of the EF97 receiver parts. The two cathode resistors have to be 1k2, no 120 Ohm. Sorry for this error.

Also, I fogot to put a value at the capacitor after the detector, C19. That has to be 47 pF

Kind regards

Nico

 

To thank the Author because you find the post helpful or well done.

Dear Nicolaas

I have seen your article about the low voltage shortwave radio. It would be a nice project during winter time. Do you have a schematic diagram about the radio?

 

with best regards from Switzerland

Aldo Patriarca

HB9HEA

To thank the Author because you find the post helpful or well done.