Very Low Noise Lentek Battery Power Supply - DIY

Very Low Noise Lentek Battery Power Supply – DIY

Finished Power Supply

Introduction

So after decades of never knowing whether my Lentek head amplifier alkaline PP9 DC battery was OK I decide to replace it with a NiCd source that I could monitor its voltage and charge and discharge as required.

It took quite some time to put this power supply together, so was it worth it and did it make ANY worthwhile audible improvements? Yes, read on.

It is generally understood that the electrochemical noise voltage from batteries will generally be much lower than that generated by most low noise analog power supplies. It is true to say that regulator devices like the LT3042/45 can produce ultra low noise levels (<1.0µVRMS at 10Hz to 100kHz), but they are still indirectly attached to mains and have the potential for injecting hum and other noise. So I decided to design a rechargeable battery power supply that when not being charged or discharged had no connections at all to AC power or AC ground. There are several types of rechargeable cells the most common being:

Lead acid
Lithium-Ion – Li-ion
Nickel Metal Hydride – NiMH
Nickel Cadmium – NiCd

Based on several white papers and tests carried out by others, see references below, NiCd cells appear to show the lowest levels of electrical noise when under load due to their very low internal resistance. Generally the higher its mAh capacity the lower its internal resistance. They are however susceptible to misuse and a shortened life span if incorrectly charged and discharged, as opposed to NiMh batteries that are far more robust, do not need to be discharged before charging, and are generally far more forgiving, providing better longevity. Despite the NiCd disadvantages I wanted the lowest noise source possible, so NiCd it was to be.

My Lentek head amplifier DC supply needs are very minimal, 9V at less than 1mA. NiCd batteries have a nominal cell voltage of 1.2V. For eight cells this will vary from about 11.2V at full charge, to 8.00V when fully discharged. So my power supply design was based around eight NiCd batteries (nominally 9.6V – 8cells*1.2V) and could support charging voltages up to 18V at constant currents up to 1 amp.

I shall not bore you with the design details, schematic diagram and a component list, but I have included a simplified block diagram of the final design below, and a brief explanation of its operation:

Power Supply System Block Diagram

Operation

The charge circuit on the main board provides a FET regulated and adjustable 18VDC supply that is set to 14.5VDC. This feeds an adjustable 1 amp constant current source set at 0.11A that charges the EBL 1100mAh NiCd’s and disconnects itself once the battery voltage reaches 11.30 volts (8cells*1.41V). The discharge circuit (the piggy back board) can be adjusted to discharge at up to 1 amp. Its fixed resistive load is currently set to discharge at approximately 0.17A. These cycles are initiated from the two front panel buttons shown below which are electrically interlocked so only one can be operating at any one time. Charging or discharging can be initiated any time. However, after the discharge cycle is entered it will automatically switch to charging when the battery voltage falls below 8.2V. Once the battery voltage rises to 11.30V the supply stops charging and disconnects the battery from the electronics, returning it to service.

During the charge or discharge cycles, the battery output terminals on the rear panel are completely disconnected from the battery and electronics by the two relays.

When powering the Lentek, the power supply is manually switched off using the rear AC mains switch.

You will notice three low noise DC fans. There are a lot of electronics in this small box and additional forced air circulation is required to remove the heat from the regulators, the discharge load and to ensure the batteries stay cool during charge. Only the front panel fan and rear bottom fan are on when powered up and/or charging. The third fan above the batteries only comes on during discharging.

The three position front panel switch allows the battery status to be monitored:

UP – During charging – voltage and current, not charging – charge set voltage
CENTER – Display off and disconnected from battery
DOWN – During discharging – voltage and current, not discharging – actual battery voltage

Fully assembled power supply

Main power supply. Voltage and current regulation & charge control

Discharge control piggyback board

Set Charging Voltage

Battery Voltage Check

Charging Voltage & Current

Discharging Voltage & Current

Electrical Noise Measurements:

All measurements: Leader LMV-185A AC Millivoltmeter BW: 5Hz-1MHz, Tektronix T935A oscilloscope BW: DC-35MHz.

Battery Power Supply:

Power supply plugged in and powered down, battery noise voltage too low to measure up to 0.5 amp.
Fully charged off-load battery voltage 10.90 volts = 1.36 volts per battery.
Power supply noise 0.02mV (20.0µV) RMS @ 14.5VDC using a resistive 0.5 amp load: -117dB relative to 14.5 volts. No DC motors. (Measured at charging circuit output)
Charging noise 0.05mV (50.0µV) RMS @ 0.11 amp: -107dB relative to 11.30 volts. Increase in noise from two DC motors. (Measured on battery terminals)
Discharge noise 0.1mV (100.0µV) RMS @ 0.16 amp: -99dB relative to 9 volts. Increase in noise from third DC motor. (Measured on battery terminals)

Added Voltage Regulation

The Lentek head amplifier is designed to operate on 9volts. While it is highly unlikely that a supply voltage range of 10.9v to 9.0v would produce any audible changes I decided that I was not going to take a chance. So I added a precision, adjustable, low noise, TL431 shunt regulator in the Lentek PP9 battery compartment to maintain the supply at exactly 9.0 volts.

PP9/Shunt Regulator Size Comparison

With a metal oxide resistive load of 10K drawing approximately 0.9mA, noise measurements on its output showed no levels that my equipment could measure, being lower than the minute level detected from the Alkaline PP9 battery.

Lentek Head Amplifier With The 9 Volt Shunt Regulator In Place And NiCd DC Feed

So were there any audible improvements? Yes, read on.

Lentek Head Amplifier Noise Measurements:

All measurements taken with the phono inputs loaded with 1.8ohms and output loaded with 47K. They include the background residual noise of the measuring system, cabling and head amp when not powered up of 0.01mV (10.0µV) RMS. This noise looked to be mostly broadband RF pickup.

Lentek Head Amplifier Left & Right O/P Noise Levels:

PP9 battery:
- 0.01mV (10.0µV) RMS (Lentek not powered)
- 0.025mV (25.0µV) RMS (Lentek powered)
NiCd battery power supply – with or without the shunt regulator:
- 0.01mV(10.0µV) RMS (Lentek not powered)
- 0.02mV (20.0µV) RMS (Lentek powered).

The above figures indicate a potential reduction of 5.0µV or a 2-3dB drop in the noise floor output on both channels when powered by the low noise NiCd batteries with or without the shunt regulator.

Comments & Listening Observations

The DC fan motors, just like the display, are digital switching devices and create a significant amount of digital noise. This cannot be easily eliminated in a DIY design like this using strip board in such a confined space. Fortunately it has no impact on the battery during the charge/discharge cycles or its electrochemical noise performance.

Clearly there was some measurable, and repeatable, improvement in the amplifiers noise performance using the NiCd batteries and shunt regulator when compared to the alkaline PP9 battery. Auditioning the supply with the master volume control at its maximum of +18dB showed both a small reduction in background noise and a change in its character. It was softer with less of an “edge”. Compared to using the PP9, no increase in hum level was noted. However, GREAT care had to be taken to ensure that neither the power supply chassis, nor the DC power cable to the Lentek head amplifier, were close to other active power supplies or AC power cables. If they were they would pickup 60Hz hum either through inductive or capacitive coupling and inject it into the DC feed. This would create an extremely low, but audible level of hum.

To be honest, powering the Lentek head amplifier from this very low noise battery supply initially provided no audible changes to most of my test vinyl, except a feeling that the stereo depth and imaging had improved slightly. That could have been psychological. However, one album in particular, Sheffield Lab 20, The Sheffield TRACK Record, revealed a significant increase in extreme bottom end power, kick drum slam, bottom end dynamics and bottom end solidity. To say I was astonished is quiet an understatement, especially as I had not noted any obvious improvements with any of my other reference test albums, all of which I know intimately but do not have the same bottom end performance as LAB 20. I immediately auditioned several more direct cuts that have a similar bottom end to LAB 20 and yes, ALL of them exhibited a similar improvement in very low end dynamics and response.

At my listening position and normal volume setting of -10dB to -6dB my system noise is already almost inaudible, so the very small improvement in noise level could not be heard.

I cannot immediately come up with a valid technical reason as to why the extreme bottom end response improved in the manner it did. Maybe the impedance of the new supply system is so much lower than that of the PP9 that it improves current delivery and voltage drop. (Any insights from readers would be welcomed on this point.)

A taxing ‘shoehorn’ build and I am not too sure that I can recommended this DIY exercise as a way to improve the overall quality of your audio. However, the improvement in the very low end of several of my records makes its continued use mandatory for me.

I cannot stress enough that the box isolation & grounding, its AC power connection, the optimum placement of the power supply and the routing of the DC cable are all important in ensuring no increase in hum level when compared to the PP9.

At least now I will not have to worry about the state of my PP9 battery any more! Also I suspect that the NiCd’s natural internal discharge will swamp the power consumption of the Lentek, so hopefully it will be many many months before I need to repeat the discharge/charge cycle.

References:

From Vinyl To Plastic

One man's quest for audio perfection

One man's quest for audio perfection

Very Low Noise Lentek Battery Power Supply – DIY