Buy Boards & Parts:
www.beezar.com
DIYForums.org
The Millett Hybrid MOSFET-MAXed - Tweaks

Millett MOSFET-MAX Relay-delay

NOTE: The Millett MOSFET-MAX (based on the MAX V1.2 PCB) takes a departure in previous MAX/MiniMAX designs. The headphone delay-relay is entirely re-designed by Toole Audio. It is a power-on/power-off delay muting and DC offset protection circuit, but with a much smaller parts count than any before. Moreover, it presents a much faster response time in the actual relay switching. Many conventional relay-delay circuits still present a noticeable thump when switching the signal circuit into the amp. The Millett MOSFET-MAX relay is much faster and nothing but a small click will be heard in most headphones. Moreover, the entire relay circuit can be easily applied to existing older Millett MAXes (V1 PCB) and MiniMAXes with only the updated parts selection detailed here and on the MOSFET-MAX BOM (more on that later).

The new MOSFET-MAX relay-delay schematic is shown above. This circuit forms a protective delay sequence that prevents signal power from reaching the headphone jack until the Millett tube circuit has had a chance to warm up and charge the capacitors. Tubes freely pass DC. As with the original Millett Hybrid, the MOSFET-MAX incorporates coupling caps (CA7-L/R) on the output to block DC. However, these caps take a number of seconds to fully charge, a prerequisite to their blocking ability. If your headphones are plugged in before the caps have charged (allowing them to block the DC), your headphones will be exposed to many volts of DC. In the case of the MAX/MiniMAX/MOSFET-MAX with V+ set at 27VDC, as much as 13.5VDC is on the headphone jack before the caps have charged. The relay circuit delays switching the headphone jack into the amplifier output until the caps have charge.

The actual time needed for the output capacitors to charge can be highly variable, but most of us that have measured it normally see about 5-10 seconds. With the parts' values shown, the relay's circuit delay has been set for this amount of time plus a gracious safety factor - about 45 seconds from a cold start. This may seem like an outlandish large safety factor (compared to 5-10 seconds). However, if the amp is turned off and turned back on before the relay cap (CM3) has had time to completely discharge, the delay will be much shorter - perhaps half the time or less. Consequently, a large delay time safety factor is desired in order to protect headphones from those on/off bumps and other spikes/sags that might occur. Tests from hot starts (bumping power on and off) have shown that the relay will still delay for up to 20 seconds. We feel this gives the best combination of protection without causing too much frustration in having to wait for the headphone jack to switch-in. After all, the tubes may take about 30 seconds to warm up, anyway.

The single most important change in the Toole Audio 24VDC Relay-Delay circuit is that the relay has been increased to 24VDC and there is a true, 24VDC regulator included in the circuit (instead of a Zener-resistor-transistor arrangement). This has a number of consequences - almost all of them good. First, the Zener diode and its associated resistor goes away completely. Second, the fact that a higher-voltage regulator is used and causes much less heat generation, it allows us to remove the large TO-220 transistor (typ. BD-139) from the circuit, too. Meanwhile, the traditional CM2 cap used to bypass the CM1 cap (the power cap for the voltage regulation) was found unnecessary when the MiniMAX was designed. However, with the 24VDC Relay-Delay, it's best to go back to the faster, higher-frequency ceramic cap. CM1 thus becomes a 0.1uf X7R cap. Finally, there is always a small bit of arcing that occurs in the contacts of a relay (without suppression). The higher the voltage, the worse this arcing becomes. This arcing can cause the relay to degrade over a long period of time. A voltage suppressor can be used instead of the traditional diode (used to be 1N4148) to minimize this arcing.

The use of the 24VDC relay results in a much faster switching time. We're not talking about shortening the delay - that is set in the circuit. Instead, the speed at which the actual contacts switch closed is faster. In traditional 12VDC relay circuits, a thump can still be heard/felt in the headphones when the relay kicks in. This is all but gone for most headphones with the Toole Audio 24VDC Relay-Delay circuit.

The following parts are required:

Headphone RELAY-DELAY
MOSFET-MAX Part #
Description Qty Common Part #
RM1 2M ohm 1 V-D RN55, 2M
CM1 0.1uF X7R 1 (same as description)
CM3 470uF 35V Electrolytic (10mm dia. Max) 1 (same as description)
DM2 SA48C (or 1N4148 if necessary) 1 Transient Voltage Suppressor
VRM1 24VDC Linear Regulator, TO-92 1 78L24 (TO-92)
QM2 MPSA14 1 Many substitutes,
check data sheets below
RELAY Omron 24V DPDT Relay G6A-234P 1 Many substitutes,
check data sheets below


Here are the data sheets for the parts in the Toole Audio 24VDC Relay-Delay. Note that we have not tested the relays listed as alternates to the MOSFET-MAX BOM's listed relay. Their ratings and pin dimensions should fit, but we have no idea if the contact switching is as fast. We have not noted any shortages in stock of the G6A series relays in several years:
OMRON G6A-234P-ST-US-DC24
(listed on the MOSFET-MAX BOM)
OMRON G5V-2-H1-DC24
Fujitsu-Takamisawa RA-24W-K Fujitsu-Takamisawa RY-24W-K
Transistor Data Sheets:
VRM1: 78L24 (TO-92) QM2: MPSA14
The MPSA14 is a Darlington transistor with an extremely high collector current - 1A or over. Accepted alternates include: KSP13, KSP14, KSP25, KSP26, KSP27, MPSA13, MPSA27, MPSA28, MPSA29, 2N6426, 2N6427, 2N7052, 2N7053, BC372 or BC373. One of the 2N5087/88's WILL NOT WORK.

The product of the values of RM1 and CM3 determine the delay time. RM1 is sized at 2Mohm, this makes the resistor a bit more difficult to find in stock at some places or at a regular low priced for common V-D resistors, but you only need one of them and it's a small price to pay for the advantages of the new circuit. The reason this was done and perhaps the only negative consequence of the 24VDC circuit is that the CM3 capacitor now has to be rated at 35V. It has to see the full voltage of the 24VDC circuit plus some safety factor. This means that the CM3 capacitor is larger at all uf ratings. So, we arrived at the 2M resistor in order to keep the size down of CM3. This kept the footprint on the MAX V1.2 PCB at the same size and it also allows the easy retrofit to any MiniMAX with this parts selection.

As before, we are recommending that you size CM3 to 470uf. As mentioned above, the 24VDC nature of the circuit effectively cuts the delay in half. However, we chose to double the value of the RM1 resistor. This will still result in a cold start delay time of ~45 seconds, just as its been with the MiniMAX. Even a fast turn-off/turn-on should still result in a delay of ~20 seconds. This is what we recommend as a minimum. The board's pads can accept a 10mm dia. capacitor. So, whatever you can fit in there will work - a high-quality cap is not strictly necessary. Moreover, many lower grade caps are smaller in size than the typical UPW or FM caps most of us use, so finding significantly higher-rated capacitors should not be an issue. 3M resistors are also available, if you find that necessary. Note also that sone previous types of relay-delay circuits are configured with bi-polar caps. Those are not necessary for the MOSFET-MAX relay-delay circuit.

Note that while the quality of the electrolytics in the delay circuit are not critical, they should be rated at the 105deg. C. rating. This is due to their proximity to the front plate and to the middle two heat sinks in the output buffer. Quite a bit of heat can develop in this section and the last thing you may want to fail is the delay capacitor.

In addition, a DIP-16 socket is a nice touch. The relay will plug directly into one and allow easy replacement. You will have to snip some of the legs on the socket where there are no pads on the board, but as with opamps, it will prevent exposing the sensitive coils inside the relay to too much heat when soldering. If you retrofit a MiniMAX to this circuit, we highly recommend that you go back with the DIP socket (two DIP-8 sockets will work quite nicely).

Here is an example of using the Toole Audio 24VDC Relay-Delay in an existing MiniMAX:
Note that the CM1 and CM3 caps used in the MiniMAX are retained. This is not optimum, but you can easily see how the circuit works and the improvement in "thump" is noticeable. Likewise, the DM2 diode remains the 1N4148. This is OK, too - unless you want the very longest life of the relay.
The important changes are as follows:
  1. Delete RM1, DM1, and QM1.
  2. Solder the 78L24 (TO-92) regulator, face forward into the pads for QM1:
    1. The left pin goes in the left hole.
    2. The center pin goes into the left hole for DM1.
    3. The right pin goes into the middle hole for QM1.
    4. Replace the relay with a 24VDC rated relay.
Simple!

file last changed:Thursday, July 4, 2013 7:00:00 AM
Please contact the MAX webmaster for questions about these web pages.