diff --git a/src/emsdevice.cpp b/src/emsdevice.cpp
index 26c7a75cd..8c5db0653 100644
--- a/src/emsdevice.cpp
+++ b/src/emsdevice.cpp
@@ -311,7 +311,7 @@ std::string EMSdevice::telegram_type_name(std::shared_ptr<const Telegram> telegr
     }
 
     for (const auto & tf : telegram_functions_) {
-        if ((tf.telegram_type_id_ == telegram->type_id) && ((telegram->type_id & 0xFFF0) != 0xF0)) {
+        if ((tf.telegram_type_id_ == telegram->type_id) && (telegram->type_id != 0xFF)) {
             return uuid::read_flash_string(tf.telegram_type_name_);
         }
     }
diff --git a/src/telegram.cpp b/src/telegram.cpp
index 6af4842e7..27de14f8f 100644
--- a/src/telegram.cpp
+++ b/src/telegram.cpp
@@ -171,45 +171,22 @@ void RxService::add(uint8_t * data, uint8_t length) {
     // EMS 1 has type_id always in data[2], if it gets a ems+ inquiry it will reply with FF but short length
     // i.e. sending 0B A1 FF 00 01 D8 20 CRC to a MM10 Mixer (ems1.0), the reply is 21 0B FF 00 CRC
     // see: https://github.com/proddy/EMS-ESP/issues/380#issuecomment-633663007
-    if (data[2] < 0xF0 || length < 6) {
+    if (data[2] != 0xFF || length < 6) {
         // EMS 1.0
+        // also handle F7, F9 as EMS 1.0, see https://github.com/proddy/EMS-ESP/issues/109#issuecomment-492781044
         type_id        = data[2];
         message_data   = data + 4;
         message_length = length - 5;
     } else if (data[1] & 0x80) {
         // EMS 2.0 read request
-        uint8_t shift = 0; // default when data[2] is 0xFF
-        if (data[2] != 0xFF) {
-            // its F9 or F7, re-calculate shift. If the 6th byte is 0xFF then telegram is 1 byte longer
-            // if not, we dont know how to handle it
-            // maybe a 24 bit type-id? Example: 90 B0 F7 00 02 02 00 00 96
-            if (data[5] == 0xFF) {
-                shift = 1;
-            } else {
-                LOG_DEBUG(F("unhandled rx: %s"), Helpers::data_to_hex(data, length).c_str());
-                return;
-            }
-        }
-        type_id        = (data[5 + shift] << 8) + data[6 + shift] + 256;
+        type_id        = (data[5] << 8) + data[6] + 256;
         message_data   = data + 4; // only data is the requested length
         message_length = 1;
     } else {
         // EMS 2.0 / EMS+
-        uint8_t shift = 0; // default when data[2] is 0xFF
-        if (data[2] != 0xFF) {
-            // its F9 or F7, re-calculate shift. If the 5th byte is 0xFF then telegram is 1 byte longer
-            // if not, we dont know how to handle it
-            // maybe a 24 bit type-id? Example: B0 10 F7 00 02 00 00 1F 00 8F
-            if (data[4] == 0xFF) {
-                shift = 1;
-            } else {
-                LOG_DEBUG(F("unhandled rx: %s"), Helpers::data_to_hex(data, length).c_str());
-                return;
-            }
-        }
-        type_id        = (data[4 + shift] << 8) + data[5 + shift] + 256;
-        message_data   = data + 6 + shift;
-        message_length = length - 7 - shift;
+        type_id        = (data[4] << 8) + data[5] + 256;
+        message_data   = data + 6;
+        message_length = length - 7;
     }
 
     // if we're watching and "raw" print out actual telegram as bytes to the console
@@ -463,21 +440,20 @@ void TxService::add(uint8_t operation, const uint8_t * data, const uint8_t lengt
 
     // work out depending on the type, where the data message block starts and the message length
     // same logic as in RxService::add(), but adjusted for no appended CRC
-    if (data[2] < 0xF0) {
+    if (data[2] != 0xFF) {
         // EMS 1.0
         type_id        = data[2];
         message_data   = data + 4;
         message_length = length - 4;
+    } else if (data[1] & 0x80) {
+        type_id        = (data[5] << 8) + data[6] + 256;
+        message_data   = data + 4;
+        message_length = 1;
     } else {
         // EMS 2.0 / EMS+
-        uint8_t shift = 0; // default when data[2] is 0xFF
-        if (data[2] != 0xFF) {
-            // its F9 or F7, re-calculate shift. If the 5th byte is not 0xFF then telegram is 1 byte longer
-            shift = (data[4] != 0xFF) ? 2 : 1;
-        }
-        type_id        = (data[4 + shift] << 8) + data[5 + shift] + 256;
-        message_data   = data + 6 + shift;
-        message_length = length - 6 - shift;
+        type_id        = (data[4] << 8) + data[5] + 256;
+        message_data   = data + 6;
+        message_length = length - 6;
     }
 
     // if we don't have a type_id or empty data block, exit