Batteries Not Included.

Though many kinds of Batteries and cells are still made quite a few shops in Limerick, Ireland only sell AA and AAA cells for Radio. Some places have C, D and PP3. The 996 (4 x "F" cell inside) has recently appeared in volume (is Eveready clearing them out?). PP9 is rare. The 1289 / 3R12 (4.5V for 3 x "B"/U10 inside) has been spotted in Castlebar. Other PPx batteries are unobtainable. Of course the battery packs for valve (tube) radios have not been available for about 40 years.

When repair and restoring Radio sets we want to be Authentic, especially on rarer sets, but who knows which sets are regarded rare and important in 10 to 20 years time?

Unlike a Painting or sculpture, a Radio is functional. A major aspect is what is it like to use and how does it sound, not just the style of the case.

 

Authentic Batteries

So it seems strange to run a portable Battery Only set on a mains adaptor/ battery eliminator if the set wasn't used like that. Without batteries it's not really portable.

Kinds of Authenticity

1. Correct socket/connector
2. Correct size
3. Reproduction case graphics
4. Real original carton/ case refilled
5. Near Original Running time
6. Similar weight. What is it like to carry about  19lb (8.5 kg) of Radio and Batteries?
7. Original type of cells completely in reproduction case
8. Original type of cells completely in a Real original case

1939 advert for an "all dry" (i.e. no Lead Acid). "Only 18¹/₂ lbs weight"

To make a "reproduction" or "ersatz" battery we need to understand some background.

 

Era of Radio Set

 1921 to late 1930s: May be battery because there is no mains electricity. Filaments may be 2, 4 or 6.3 or 12.6V using Lead Acid cells (A Cells, HT can be 45 to 145V using Lead Acid or "Dry cells" (Zinc Carbon) (B Cells). Often HT is still called B+ in USA.  Usually there is a Zinc Carbon "grid bias" battery (C cells), though some sets (even mains models) used Mallory Mercury cells. Early Bright Emitters used a lot of power and not portable.

 Late 1930s to 1938: The Portable models used Octal 2V filaments at typically 100mA. Late models used 2V 20AH "Gel" cells, a primitive version of today's sealed Lead Acid batteries in Alarms, emergency lights and small UPS systems. The most portable 4 tube sets use 800mW

 1938 to 1953: Octal (1938 Sylvania), Edge connect (1939 Mullard/Philips), B7G (1940 RCA) 1R5, 1T4, Rimlock B8A (1948, e.g. DK40). 1.4V at 50mA apart from audio output which is 1.4V at 100mA or 2.8V at 50mA for series use. Some sets very portable, others large models for those with no mains Electricity. Filament power for a whole radio is 350mW. HT is 400mW to 900mW. Some sets use Combination "balanced" Batteries in a single package for LT and HT.

 1953 to 1960: Philips develop 7 pin mini tubes (B7G) with only 25mA current. A portable can now use as little as 175mW for Filament and less than 500mW HT. Some large Table models not intended to be portable are still produced. Perhaps 500 hours on a B137 pack for an Ever Ready Sky Prince (1954 to 1957). Some sets use Combination "balanced" Batteries in a single package for LT and HT.

Post WWII German Portable sets (and Philips) tend to add a DEAC (a NiCd rechargeable cell) and even a slight trickle charge to "maintain" the Zinc Carbon HT pack. The DEAC also acts as voltage regulator for mains operation.

 1955 to now: Transistor sets with a very wide profusion of batteries. Some still take two sets! The Sony ICF2001D uses 3 x "D" cells for the Radio and some "AA" cells for the clock/microprocessor. Some 2012 small sets with DAB use up to 5x power of 1980s models. Earlier sets used PP type packs or 2 x 3R12 as well as AA, AAA, button cells, coin cells, C and D cells. A few sets used a "duplex" 3V cylindrical battery.

Orignally almost all the Batteries were Zinc Carbon. One advantage of Zinc Carbon today is that the self discharge may be lower. Some few sets used Mercury button cells for grid bias and NiCd Rechargeable for LT.

Now we have NiMH, Lithium primary (two main kinds) (coin cells and others), Lithium rechargeable, Lead Acid Gel cells/Batteries and for 1.4V to 1.5V nominal primary cells we have Zinc Air, Silver Oxide, Alkaline, Lithium Iron Sulphide, Zinc carbon and Zinc Chloride.

Technical people immediately think of Rechargeable cells and Inverters for HT (not very Authentic, poor shelf life, cell balancing issues on HT packs, RF interference etc).

Aspects to consider

• Safety
• Shelf life for Intermittent use
• Run time
• Cost
• Authenticity

Also 1.5V and 90V batteries never actually existed, look at the battery discharge curves! The filaments are designed for a nominal 1.4V (slightly less than this is advisable on mains PSU) or in some cases 1.2V (filaments especially for NiCd).

Alkaline and Zinc Carbon AA cells with 43 Ohm load.

When designing battery packs and studying Battery mAH specs (the capacity) we need to understand the load (the Radio) and the Battery Manufacturer.

For our application the Alkaline DO NOT last five times longer, but about twice as long! The Lithium Iron Sulphide (lasts twice as long as leading Alkaline Brands) is no better than Alkaline. It's often claimed NiMH last twice as long as Alkaline. That's only true at high currents. It might be true on a 1926 valve set. But on a 1954 Valve portable the Alkaline cells could last twice as long as NiMH!

Two important figures for Batteries or cells are Internal resistance and Self Discharge.

 

Self Discharge

The NiCd are far worse than Zinc Carbon. NiMH is worse still apart from "Eneloop" type, NiMH can be as poor as 2 weeks (2700mAH AA) or nearly 2 months (1800mAH AA) or over 1 year (Eneloop approx 2000mAH). A Zinc Carbon can be 2 to 5 years depending on temperature, type and quality. Alkaline can be over 5 years, even nearly 10 in cool conditions and should not leak in storage so easily. Lithium primary cells are about 3V (except for Lithium Iron Sulphide family) and can last 5 to 20 years.

Mercury Cells are illegal and for Grid Batteries need to be replaced with Silver Oxide, Alkaline Maganese or Lithium primary cells to give the voltage required. Zinc Air button cells are for continious use once activated (hearing aids etc) and don't have low enough self discharge for a grid battery.

Internal resistance

This is lowest for NiCd and NiMH. It's so low that they need a fuse as a fault can set wiring on fire. The Lithium Iron Sulphide is of no advantage as the low resistance (which gives better performance in high currrent drain devices) isn't needed. The valves that need high current really need to be run from Lead Acid cells, as they were originally, the NiFeS, NiMH and NiCd are the wrong voltage. They are low resistance so care must be taken. Alkaline will perform about twice as well as Zinc Carbon using D cells for 125mA filament sets. But due to lower internal resistance than Zinc Carbon, they will perform a bit better  than twice as well on 250mA filament sets, where the Zinc Carbon is a used a little more than twice as fast.

Capacity & Voltage

Unfortunately the IEC standard battery tests are a different load resistance for AAA, AA, C, D and PP3 types. These loads don't match our valve radios, and often not recreated PP packs for older transistor sets. The capacity curves assume 0.9V per cell end point. This may or may not work for an HT pack, but it's too low for sensible filament voltage. However we can measure the effective internal resistance at the load conditions of our Model. The internal resistance does rise though as the battery is discharged.

The initial voltage with no load of Zinc Carbon is nearly 1.7V. A 60 cell HT pack (or 10 x PP3) for 90V will be about 101V with no load when fresh and 54V is the theoretical end point. You can "tap" the HT line to the fresh battery pack to find the "reasonable"  end voltage B+ for the radio, but do it with a 1.1V alkaline D cell as LT (or Lead Acid cell / PSU down at 1.84V) for a realistic answer.

A fresh Alkaline cell is about 1.58V, nearly 1.6V. A 60 Cell HT pack for 90V ought to read about 96 to 97V with no load.

Neither the filament nor the HT line is a linear resistor with voltage, unlike the IEC battery tests.

The 2 hours per day gives longer time on cells with reasonably low self discharge. The end voltage may be a bit low (That's 54V if 10 cells are used for a 90V battery pack. At 8V the 620 Ohms  the current is about 13mA. A set using B126 (which is about the smallest 90V pack) is likely about 8.5mA  HT. So then instead of 25 hours to reach 60V, the time would be 38 hours. But due to the Internal resistance the time is not linear, at lower currents you get more capacity.

Here is the Alkaline PP3 on the same scale. Because it's AAAA (or similar) cylinders it doesn't fill the box as much as Layer cells.

That's about 40 hours with a 620 Ohm load (for 6V end point), not even twice the 25 hours. At 8.5mA  we get maybe 61 hours if it was linear

In practice the 10 off ZnC PP3 based battery pack may give nearly 45 hours or as little as 35 hours depending on the model. The improvement  with Alkaline isn't x5 as suggested in advertising, it's x1.6, yet price is typically twice to seven times! We might get 57 hours to 70 hours (lower internal resistance means variation of capacity is as much with load). This means that for "authentic" run time use in a larger pack than a B126 is very poor. The B126 was a layer cell and may have given about 80 hours or more judging from difference in volume between 60x  layer cells used in a PP3 and the size of the B126 layer cells

B126 original 60 layer cells
(each stack is 22.5V nominal)

Coin Cells CR2032 for smallest HT packs?

The CR2032 are €2.50 EACH in the Supermarket. Even if only 30 needed (30 x 3V = 90V) that is very expensive. But they are about €0.12 each including postage in packs of 80 online.

Coin cell IEC test

Typically a fresh cell is about 3.2V with no load and the realistic end point is 2.75V, or at 0.2mA about 1000 hours! But at 8.5mA we don't get 15 hours. The internal resistance is high and really 2mA is the practical maximum current. Since they are small, we can parallel six stacks using 1N4148 diodes. That means at 8.5mA the current per stack is on average 1.4mA. Neglecting Internal resistance (which we can't) then the 1000 hours at 0.2mA is 143 hours. In practice much less.

How many do we need? Look at the END POINT, not the fresh voltage. For 60V (same as the PP3), we would only need 22 cells, but that is neglecting internal resistance. Using 26 cells gives 71.5V which is very good and 3.2mm x 26 =  83.2mm, well within internal height of a B126 less connector, insulation and terminals. With CRddzz coins the first two digits are the diameter in mm, the last pair of digits is mm thickness x 10, so CR2032 is 20mm diameter and 3.2mm thick.  With 26 cells the no load voltage is 83.2V and with 8.5mA load about 78 V. All less about 0.6V for the diodes.

This is a total of 156 coins which is €18 to €20 including postage for 160 cells. Run time could be 100 to 135 hours, or nearly  twice Alkaline:

Summary of 8.5mA replacement "90V" pack. A B126 may have been over 80hrs

The Zinc Carbon PP3 obviously give closest to original characteristic, but a lot less run time as they don't fill a B126 box efficently. On larger packs where you can fit two sets in parallel, but 60 x AA cells  don't fit, they are the best choice. A few packs can be filled with AAA or B/U10 cells from the 1289 / 3R20 flat 4.5V cycle batteries. The Alkaline PP3 may be closer to B126 capacity, but still less as the 6 x AAAA (or similar) don't fill the case well.

The coin cells are the highest "up front" cost, but the performance is almost like a mains supply, it's so little droop with usage. They are also a fire/explosion risk having so many packed together, perhaps  six pico fuses for safety? They require more work to make a holder.  They are really only a solution for the smallest packs.

What about Zinc Chloride? They are often twice to four times the price of Zinc Carbon and at best appear to offer about 50% more capacity. I'd need to research shelf life. But discount brand Alkaline are often a similar price.

D size Cell capacity as function of load. Log Log scale

This interesting graph is unfortunately a log scale for capacity. But you can see that below a particular current the capacity is close to maximum, about 8,000mAH for ZnC (though this graph from a company selling ZnCl suggests about 6,000mAH, 8,000mAH for ZnCl and over 15,000mAH for Alkaline. Actual D or F  Zinc Carbon cells are common original type for LT, either singly, in parallel (up to eightin a pack!) or series (7.5V is most common, but also 9V and 12V). Five of the B cells (U10) are also commonly used for series 7.5V LT. The F cell is same diameter as the common D cell but is almost exactly 50% longer. The lower internal resistance means more than 50% more "running time". The AD35 pack is very like two F cells in parallel, but they are slightly shorter. However if does mean we can deduce that at 125mA load it is more than 50% more than 8,000mAH x 2, probably close to 24,000mAH and on 250mA close to 20,000mAH.

For LT packs / supplies

• D cells can be replaced by modern ZnC D cells, though they won't last quite as long as most brands now double the fat plastic bung at the top that replaces tar on old ones, to reduce the amount of active ingredients.
• F cells in packs can be replaced by real F cells (128 gram) from a 996 6V lantern pack (4 cells  or the larger screw terminal 6V Lantern pack which has 8 cells), or by Alkaline D cells (145 gram).
• B /U10 cells in packs can be replaced by real B cells (33g) from a 1289 / 3R20 / 312G Cycle lamp (3 cells, two brass strip connections plat pack). These are about 2500mAH, so can't be replaced by ZnC cells, but an Alkaline AA is slightly more capacity at 2700mAH though lighter at 24 g.  
• A NiMH C cell can fit well in a DEAC 3.0 or DEAC 3.5 case made of coffee can or other material with label like original and provide similar capacity (3500mAH). It also provides the same regulation function. Avoid eneloop or higher capacity than 3500mAH as they will not like the continuous trickle charge. An actual NiCd C cell would be better, but have lower capacity. 

 

Next section will be some practical examples of using  Zinc Carbon, Alkaline and Primary Lithium. Later we will look at DEAC replacement, the less authentic rechargeable systems and inverters.

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