A Little Help For Eico's HF-81

The HF-81 is one of the very first stereo amplifiers produced when stereo first came into being. A few were sold as factory built units, but being a product of Eico, the vast majority of these units were sold and constructed as kits.

The unit went through a couple of production changes over its run, most notably adding a 32 ohm output tap and more Function Selector options along the way, but other than these few changes, the unit remained largely the same from the first to the last units to roll off the production line.

True to Eico's design philosophy, the HF-81 has a lot going for it over many of the other low priced competitors of the day. It has a full array of input and switching facilities appropriate for the day, feedback RIAA equalization, ample gain, active feedback type tone controls, and a very good set of output transformers. And, it has that EL84/6BQ5 sound that has endeared so many. Eico always did throw the majority of a unit's cost makeup towards design features in the direct audio path. As a result, the HF-81 produces a sonic result that virtually everybody appreciates. However, as with all compromises, that means that other areas of the design did not enjoy the same level of attention. Just as Eico is well known for its audio quality then, it is equally well known for the hum and noise issues that often come along for the ride.

There are multiple issues that cause this problem over the whole of the Eico line, with the hum and noise components each originating from separate design issues. Thankfully, in the HF-81 however, the noise portion of the audible quiescent signal is fairly well controlled. On the other hand, the hum level is a different matter. Collectively, the hum and noise specification from Eico for the HF-81 is a bottom of the barrel -75 db below rated output, which the original unit can just barely meet with proper attention. However, pair this specification up with some sensitive 100+ db efficient speakers in a quiet room, and there's typically enough noise available to carry on a running conversation with. In my smallish 12 X 16 listening room, this was easily audible from my listening position. When it comes to high quality audio reproduction, any notable noise that does not originate from the original program source then must detract from the sonic presentation.

In some degree of fairness, I doubt that Eico ever considered the possibility that their little intro stereo kit would ever be paired with high cost high efficiency speaker manufacturers like Klipsch and Altec. After all, this was a kit designed to save precious consumer dollars, so big efficient speakers were virtually never a part of that equation. Therefore, the lower efficiency that typical speakers of the day had that this unit would most likely be paired up with would run interference to cover up a good bit of the noise. Today however, that scenario is often blown out of the water, where high efficiency speakers are virtually the norm now with lower powered amplifiers. In that setting, the hum issues the HF-81 become very apparent rather quickly.

The vast majority of the hum component can always be traced back to a minimalist power supply effort. And minimalist it is. Compared to the efforts that more up scale manufactures used to control it, the Eico has but a lowly single hum control to battle it, which only tames the most gross amounts of the stuff. But there is help that can make a solid improvement in the hum levels produced -- one of which also has notable other benefits as well.

So, this thread doesn't set out to retrace the excellent efforts of others before me regarding this amplifier, but rather, offers some fresh new input on areas I have not seen addressed with this unit. Therefore, the ideas presented can be applied to those units that have already had a good thorough makeover.

Stay tuned --

Dave

First off, Great post Dave! I have been using an HF-81 amp in my garage and I did notice the hum when all is quiet. My speakers are B&W and I believe about 94 db Efficiency. I tried tuning it out using the hum pot but it would not get quiet. I did a small mod, I found a 500 ohm 2W pot that would fit there and installed that similar to the original design except I did not ground the wiper, instead I hooked up a 180k resistor from the last capacitor in the filter stage and connected it to ground through a 20k resistor. The junction of these two points I connected a 22uF 50V capacitor to ground. This junction between the resistors also goes to wiper of the pot. This mod supplies about 20 VDC to the filaments and provides a bleeder for the caps when the amp is off. The pot works much better to null the hum, and I didn't have to drill or mount anything to the chassis! Other ways I was thinking about was to rectify the 6.3 and feed it to the 12AU7's and the 12AX7's in the amp. That would have been more work and parts.

I have 1 of these an still have done nothing with it..
(may need to dig it out)

I should have bought one before this thread started ... :tears:

I always appreciate all your threads as there's always something to be learned whether we have the subject pc of gear or not.

Tuning in for the tutorial... Thanks for sharing.

I've recently returned to my HF-85 project, so the long-delayed PEC replacement board is back on the design table.

This is one of those things that measure best with the human ear. I like rebuilding the front end as designed, but with parts instead of PECs, then you can play with the output stage and power supply all you want, and it will still sound good. Last one I did got SS regulated screens, and DC on the phono fils. The phono stage doesn't need much at all to sound very good compared to the competition IMHO.

First Steps and Identifying the Noise Sources

Let me again emphasize that this thread is not at all so much about the sonic characteristics of the HF-81, but rather, about its social conduct in polite society -- and more specifically, the amount of hum it contributes in that setting.

There are three principle noise sources in the HF-81, all of which require their own separate approach for resolve. These include:

1. The marginal B+ filtering as applied to the output stage,

2. The use of AC voltage to power the small signal tube heaters, and

3. The power transformer itself.

Each of these present their own unique contribution of hum timber to the total hum produced, and will be taken in the order presented, with a few pics along the way to show how they were addressed in the subject HF-81 at hand.

B+ FILTERING:

Because the B+ for the plates of the output tubes is sourced directly from the cathodes of the rectifier tubes, there is a very limited amount of filtering achieved before this power is applied to these elements of the tubes. As is typical of so many small amplifiers, Eico then depends on a good DC balance in the output stages to reject the most of this noise. This approach works pretty well, but requires well matched output tubes to really get the job done.

From a circuit standpoint, the usual approach to reducing hum from this source is to up the amount of filter capacitance, but since the filter cap of concern is also connected directly to the rectifier tubes, that increase can only be made modestly, or the rectifier tubes will become toast if any hot switching events take place. So while any increase in capacitance will help the hum issue, it can only be of modest help.

To effectively deal with this hum source then really requires a two pronged attack by:

1. Reducing the noise at the source, and

2. Reducing output stage sensitivity to the noise.

Such an approach will not just reduce, but effectively eliminate this noise source from the output signal.

REDUCING POWER SUPPLY NOISE

Reducing power supply noise can most easily and effectively be accomplished by adding a choke to the filter network before the output stages connect to this supply. But space is limited in this little guy, with no possibilities presented under the chassis. On the top side, there's really no good possibilities here either if a neat installation is to be made of things. Therefore, in the classic art of compromise, something's gotta go.

In this case, the most logical thing to go is the rectifier tubes -- and this presents some real benefits:

1. The heat source from the tubes is eliminated.

2. Nearly 13 watts of energy is no longer required from the power transformer.

3. Rectifier tube replacement caused by a variety of reasons is eliminated.

4. Space is freed up.

5. Other possibilities become viable.

So the rectifier tubes go, and if proper consideration is given to the installation of new SS rectifiers, then you'll never miss them. That includes:

1. Making sure that all filter caps throughout the unit can handle the (typically) 425 volt surge produced at turn on when SS rectifiers are used. If a recap is being performed, then use of new 500 or 525 volt caps through out eliminates this concern.

2. Installing an appropriate series dropping resistance BETWEEN the output of the SS rectifiers, and the first filter cap. This is a hugely important point that is so often missed when SS rectifiers are substituted for prior vacuum tube rectification -- IF -- the original heated characteristics of the vacuum tube rectifiers are to be maintained. Use of an appropriate resistor then not only minimizes the (new) turn on current surge that SS rectifiers produce, but also produces virtually identical operating voltages -- AND -- voltage regulation characteristics as provided by the former rectifier tubes. When installed in this fashion, the audible difference between SS and rectifier tube installations really becomes a moot point.

In applying these thoughts then, the rectifier tube sockets were drilled out (this was a factory wired unit), appropriate terminal strips were installed for mounting the new components on, a 200 ma choke was neatly mounted on the divider plate between the old rectifier tubes and the tone control circuitry, and the old vacuum tube rectifiers have never been missed.

To complete the new pie filter created by the addition of the choke, an additional filter capacitor was required. Space for this was obtained by simply making the old 3 section can cap a new 4 section 525 volt can cap, providing a neat solution. The two caps over in the power supply section now operate on either side of the choke, while the filter cap for the output tube screen grids (originally one of the two in the power supply area) now resides inside the new can cap.

Along the way, the hum balance control was also relocated over to the outside most rectifier tube socket location, freeing up needed space for other circuit modifications that were implemented to be discussed later. A steel washer was mounted (via JB Weld) under the old tube socket hole to effect mounting of the control in its new location.

In view of the total modifications done, this portion of them has solved a number of problems, without introducing any new ones. The amplifier and power transformer runs cooler, rectifier tube replacement is a thing of the past, other modification possibilities have been created, and all noise from 120 Hz power supply ripple has been eliminated -- whether matched output tubes are installed, or not.

A few pics show the results of this work:

1. Gone are the old rectifier tubes, and in are terminal strips, SS rectifiers, and a 50 ohm 10 watt dropping resistor connected between the output of the rectifiers and the first filter cap. This resistor dissipates only 1.5 watts under normal operating conditions, which represents a small addition of heat, compared to the 13 watts of heat lost from removing the rectifier tubes. Most importantly, the nearly 6 volts average ripple voltage that was originally being sent to the output transformers has now been reduced ~ 60 times, to about 100 mv. B+ voltage and voltage regulation characteristics closely mimic that of the original rectifier tubes. Installation of a CL-90 current limiter device (seen in the shadows) rounds out the conversion from thermonic to SS rectification. The relocation of the hum balance control worked out very nicely in this area as well.

Also, two other efforts can be seen in this pic to help minimize hum:

A. The wiring to the panel lamp is now shielded. The AC voltage in this wiring runs across the top of the tone controls circuits, and is a a potential source of hum voltage for these circuits to pick up. Shielding the wiring prevents this possibility.

B. The other shielded cable services a popular modification for the HF-81: making the old Tape Speed switch the AC power switch. Even though the old power switch was still good on this unit, making this swap eliminates needless wear on the treble control for the right (amplifier #2) channel. Using shielded cable here also eliminates any new source of hum noise this modification might make.

Finally, there are three other hints of hum reduction efforts yet to be discussed in this pic. They'll be coming up soon.

2. On the top side, access for the hum balance adjustment can still easily be had, the choke fits nicely in the newly freed up space, while a grommet installed where the inner rectifier tube used to reside allows access for wiring to pass below the chassis. The additional components there will be discussed in a later post.

Next up, reducing output stage sensitivity to power supply noise.

Dave

It looks like you're running SS diodes in series, or am I mistaken? What's the benefit for that?

Squid -- At turn on, the peak inverse voltage developed across each leg of this conventional full-wave rectifier circuit is about 850 volts. While a single 1000 PIV rated diode on each side will certainly work, this leaves precious little room for any AC line spikes that might come along at an opportune time. In fact, just a 31 volt (peak) transient riding on top of a 120 volt RMS line voltage at the right time would create a peak inverse voltage that begins to exceed the 1000 volt rating of a single diode. By using two diodes in series on each side, this rating is doubled to 2000 PIV, leaving virtually no chance of the diodes ever being damaged by way of AC power line spikes. It simply helps to make the power supply bullet proof.

Thanks for the interest!

Dave

That's definitely one way to do it, and I like it. Some tube purists might buck a little though. I generally always use at least one thermistor when going from tube to SS rect.. I myself have never done a SS conversion on an 81 or 85, for the simple reason that I'm not sure why they are so pleasing to the human ear as is. There has to be something to a following that commands 4 or 5 bills for a fixer upper. I've used Heyboer's excellent replacement PT which is what the original should have been (current wise). I'd really like to hear this SS job next to an original. Don't mean to hi-jack the thread, and I hope I'm not, but I've never really had a noise problem with these using a (quasi) star. Can't wait to see more. :yes:

Good read... Most SS PS circuits I've looked at seem to start with a cap, or choke first, however using a dropping resistor 1st would put less load on the components following it... never thought of it as an option, but it looks like the way to go...
Thanks again for the effort...

Regards,
John

dcgillespie said:

Thanks for the interest!

Click to expand...

Really, it's me who should thank you for the interesting post and the great diode explanation: I'm not much of a sandman, but I'm always learning!

Sidebar: Conversion to SS Rectification

This topic always seems to bubble up whenever the conversion from vacuum tube to SS rectification is made. Many insist the change is audible, and I would hardly suggest that they hear otherwise.

Virtually always however, the comparison -- and therefore the judgement -- is never made on a level playing field. Out comes the tube(s), and in go the diodes. There are obvious and well known changes that take place when this change is made:

1. Turn on surge currents increase.

2. Heater current decreases, and

3. B+ voltage rises.

And then there is the ever present discussion of SS diode switching transients, as well.

But there are other changes that occur, because whereas the SS diode is basically a one way switch with virtually no impedance when biased on, the vacuum tube is also a switch, but introduces significant impedance (by comparison) into the circuit when forward biased. As a result:

1. The solid state "conversion" has significantly higher peak current flow occurring even under quiescent conditions -- in much the same way that a fixed bias output stage can have much greater peak current flow than a cathode biased output stage can. Why? Because of the cathode bias resistor inherently limits peak current flow through the stage. Similarly, the impedance of a vacuum tube rectifier limits peak current flow from the power transformer in much the same fashion.

2. The greater impedance of vacuum tube rectifiers reduces not only B+ voltage, but power supply regulation as well, just as inserting any resistance between the power supply and load would.

3. The vacuum tube rectifier typically results in little change in power transformer operating temperature, as while the loss of supplying any heater current is significant when converting to SS operation, so is the increased peak current demands that SS rectifiers place on the HV winding of the power transformer -- enough to virtually wipe out all the gains made, as increased heating of the HV winding is significant when a straight conversion to SS operation is made.

This is why I stress that it is so important to place an appropriate resistance between the output of the SS diodes, and the first filter cap when a conversion is made. THEN, other than the instant on characteristic of SS diodes, the combination of diode and resistor mimics the warmed operating characteristics of a vacuum tube rectifier very closely indeed. In fact, so configured, it would be hard to distinguish -- by simple characteristics -- the difference between this configuration, and the quick heating 5U4 type family of rectifier tubes.

In addition to minimizing transformer heating and mimicking tube rectifier voltage drop characteristics, the resistor also acts to greatly diminish the turn on surge that straight out conversions create -- making the need for current limiting devices -- based on their defined purpose -- greatly reduced, all else being equal when conversions are made.

Finally, with the peak currents contained, so are switching transients. The resistor and first filter cap acts as a highly effective snubber for the diodes, eliminating the switching transients just as a snubber would across any switch contacts. The impedance of a rectifier tube acts to reduce switching transients in the very same fashion.

So if the playing field is to be level when comparisons are made between SS and thermonic rectification, then the resistor must be included to in fact level the field. But I dare say that so many comparisons and ultimate judgements have been made based solely on a straight out replacement comparison. Nothing could be further from setting up a fair comparison. And certainly hopefully, the straight out comparison wasn't made by using a variac to bring the B+ voltages back in line, as then heater voltages can also become compromised creating other problems, as will be shown when we get back into the project at hand.

If you've made up your mind against SS rectification based strictly on sonic analysis, but it has been based on straight out conversions between the two, then you owe it to yourself to revisit the issue, where the SS rectifiers have been correctly installed to properly mimic the original tube devices. The look and attraction from the warm glow of the tube rectifier is undeniable, and will never be equaled by any SS device. But if you can look beyond those things from a purely sonic standpoint, then properly implemented, the sound produced between the two will likely be indistinguishable.

Dave

Just curious, I know these are infamous for killing stock PTs, but is it typically the primary, or secondary that gives it up?

A number of manufacturers have models that are known to eat power transformers: Fisher's X-100, Heath's mono power amps through the W-5M, Eico's and HF-81, 85 & 86 to name a few. Keeping them cool is the best defense against normal use failure.

Usually, it's the HV secondary that gives up. It's the winding with the finest wire, and is invariably the the last winding wound on the bobbin. Therefore, it is subject to the most expansion and contraction with every heat cycle. Once the insulation on the wire goes, shorts happen, it then runs much hotter, and that's all she wrote.

Dave

Dave,
what's holding those caps in place ? Silicone?

Silicone adhesive. It does a great job, and is unbothered by the heat of operation.

Dave

Thanks Dave this is a great read, I have one which I restored and the hum is low in
mine.The vibration from the transformer is another matter I did a lot of work on it and
got under control. I think in shipping the transformer coil shifted over touching the
metal frame which I had to shift over and center it and added wood strips to hold it in
place which seems to help. I will in time be getting a new Heybour PT for it but for
now it's running good.

Tube

The only issue I had with the Heyboer was one of the through bolts wasn't tight. It's impossible to tighten them after it's in place and wired, so it had to come all the way out again. Taught me to always check them. I use nylon washers now too, on both sides of the through bolt and on the chassis where the TX bolts down. Hardware stores have them in all different sizes and thicknesses.